Needs you

Every question only you can answer, across the handbook. Each one carries a proposal you can accept as it is.

723 open

Company and strategy 43

• The one page
  Is Fetch the direction? Proposal: decide when Param picks one of the five; until then this version is used only when someone reads in the Fetch direction.
• The one page
  Seven sources or 68? The books say seven sources tick by; the inventory lists 68 in seven groups. Proposal: say 68 sources in seven groups on the page.
• The one page
  Do we show the partner count? Proposal: say "real brands working with us now" and give no count until partners agree in writing.
• Story and why now
  Which finding leads the story? Proposal: the AI answers gap, since it is the newest shelf and the easiest to check live on a call.
• Story and why now
  Whose words replace the Samples? Proposal: ask each of the five partners for one sentence about their Monday on the next call, with written permission to quote.
• Vision and mission
  The whole document waits on you: Where this goes in five years; the sentence we hold ourselves to
• Vision and mission
  The vision. Proposal: "In five years, every online store has a marketing team that reads its store every week, drafts what to do, ships it with the owner's yes and learns what worked."
• Vision and mission
  The mission. Proposal: "Give every store owner this week's marketing, drafted from their own store, with a source on every number, and keep them in charge."
• Vision and mission
  Does the internal plan to own outcomes belong in the vision? Proposal: keep it out of the vision line and in strategy, so the vision can be shared with partners.
• Principles
  Is this order right? Proposal: accept it; truth and read only sit above everything because breaking them loses a partner's trust for good.
• Principles
  Does rule 9 cover pricing pages too? Proposal: yes; the plan and the wallet are sold on time saved, never on a percent off.
• Strategy and bets
  The whole document waits on you: Three to five bets, what proves each right or wrong, the order
• Strategy and bets
  Are these the bets? Proposal: accept bets 1 to 3 for the design partnership and keep 4 and 5 as the later order.
• Strategy and bets
  Who owns each bet? Proposal: Param owns 1 and 2, Sai owns 3 with the cost per brand work, Sai reviews 2 with partner calls, since team and roles proposes Sai for design partner relationships.
• Strategy and bets
  Review date for bets 1 to 3. Proposal: the end of the six week partner programme the book proposes, or 7 Dec 2026, six weeks from today, if no programme date is set.
• North star and goals
  The whole document waits on you: One number, the goals under it, quarterly targets
• North star and goals
  Is the north star weekly hand off stores? Proposal: yes, as defined above, counted by hand from brief records until the decision log ships.
• North star and goals
  Quarterly targets. There is no baseline. Proposal for the quarter to 31 Dec 2026: all five partners activated, and three of five handing off three weeks running; set the next quarter's targets from those four weeks.
• North star and goals
  Does a copy of the brief count as a hand off, or only a sent one? Proposal: both count, since v1 never sends on the brand's behalf.
• Market
  The whole document waits on you: Size with sources, segments, where we start
• Market
  Which market do we size first? Proposal: India, since every partner today is Indian, then the US with the same method.
• Market
  Do we pay for a store lookup? Proposal: start with free public counts and one month of a lookup tool only if Param says yes.
• Market
  Do large brands stay in the segment? Proposal: keep Puma India as a test of fit and decide after the partnership which segment we sell to first.
• Business model
  Is Plan plus Wallet plus Autopilot the model? Proposal: yes, as written on 26 Oct, with Sai's agreement recorded in the decisions log.
• Business model
  Which number sets the plan price: the $15 estimate or a measured month? Proposal: a measured month on three partners, and the plan stays at $79 only if cost sits under a third of it.
• Risks
  Who owns the monthly review? Proposal: Param, with Sai on money and partners and Arvind on platforms and integrations, as in team and roles.
• Risks
  Do we write the partner agreement now? Proposal: yes, before any partner graduates to a paid plan.
• Investor story
  The ask. Proposal: leave the slide out until Param decides whether and when to raise; see fundraising and cap table.
• Investor story
  Can we name a partner count or a partner? Proposal: "five real brands" only after each agrees in writing, names only with a signed yes.
• Investor story
  Does the plan to own outcomes reach investors? Proposal: say it in person only, never on a slide, as the brief keeps it internal.
• Investor story
  Team slide roles. Proposal: Param writes one line for each of Param, Sai and Arvind in team and roles, and the slide copies it.
• Investor update template
  The whole document waits on you: Monthly: numbers, wins, misses, asks
• Investor update template
  Who receives it? No list exists. Proposal: start with anyone who has put money or credits in, plus two advisers Param names, from the November update.
• Investor update template
  Do we report cash and burn? Proposal: yes, from the spend ledger once it exists; leave the line out until then rather than guess.
• Investor update template
  Which name heads it? Proposal: the legal entity until the product name is picked, then the product name.
• Markets and expansion
  India first, or global from day one? Proposal: India first through the design partnership, sold in English, with no India only code; the product stays global.
• Markets and expansion
  Which second market? Proposal: the US, sized with the same method as India, chosen only after a sourced count and one US partner.
• Markets and expansion
  Price in rupees or dollars for Indian partners? Proposal: rupees at the regional price, as the pricing proposal says.
• Fundraising and cap table
  The whole document waits on you: Rounds, investors, terms, the cap table, the next raise
• Fundraising and cap table
  Do we raise, and when? Proposal: not before the design partnership shows a weekly hand off habit and a measured cost per brand, so the story rests on our own numbers.
• Fundraising and cap table
  Is the LLP the right vehicle for investment? Investors in India usually expect a private limited company. Proposal: ask the CA this month whether to convert before any priced round or note.
• Fundraising and cap table
  Who owns what today? Proposal: Param writes the current split into the cap table file with the CA and links it here.
• Fundraising and cap table
  What is the first ask? Proposal: leave it blank until question 1 is answered; never put an invented amount on a slide.

Product 61

• Product overview
  Is Gully Labs still a design partner? Proposal: treat it as a past prospect and keep the five in brew-live-stores as the list.
• Product overview
  India first, or global from day one? The read asks the market and never defaults it [src: CLAUDE.md:31-32]; every partner today is Indian. Proposal: India first through the design partnership, with no India-only code.
• Product overview
  The product's future name. Proposal: decide it with the brand direction, then rename in one pass.
• Product overview
  When does v2 start? Proposal: after the v1 redesign ships to production and two partners hand off weekly for three weeks running.
• v1 PRD
  Is the Pro plan sheet still wanted on production? It shows a trial with no billing. Proposal: hide it until pricing is agreed, so no partner reads a plan we cannot sell yet.
• v1 PRD
  Do the "How we measure it" sections stay in v1 briefs? The brief writes a measure plan for the team [src: jam-brief.js:94-95], and the redesign removes measuring words. Proposal: keep it in briefs as "How your team measures it", remove it from cards and the card page.
• v1 PRD
  Seats or Which of these do you handle first? Proposal: ship the question as a filter first and keep the seat field as the ranking input until a test shows the same top five.
• v1 PRD
  The drafts cap. Code says $4 a brand a day, the loop doc says $1.50. Proposal: keep $4 and fix the doc.
• v1 PRD
  Read time. Proposal: time five reads on production this week and replace the 3 to 4 minute figure with the median.
• v2 PRD
  Which channel posts first? Proposal: Instagram and Facebook Page posts through Meta, because the Meta app and business verification are already in progress.
• v2 PRD
  Who may approve a post? Proposal: admins and members both may approve, only admins may turn the dial up.
• v2 PRD
  Which three workflows ship first? Proposal: Answer a new question about you, Get ready for a festival, Brief your agency on a heart, since each can end in a brief before posting exists.
• v2 PRD
  Does Prepare need write access? Proposal: yes, so Prepare and Schedule ship together per channel.
• v2 PRD
  v2 price. Proposal: keep $119 or ₹7,999 as the working number and decide with Sai after the partnership.
• v3 PRD
  How is a play graded fairly? Proposal: only A and B results move ranking; C and D show to the team but never teach the feed.
• v3 PRD
  When is a result ready to read? Proposal: the family's decide day from measure-families.js, with early reads shown as notes only.
• v3 PRD
  Holdouts need sending. Brew sending is cut [src: SCOPE.md:38]. Proposal: run holdouts through the brand's own Klaviyo or WhatsApp tool, never Brew's own sending.
• v3 PRD
  Keep How we measure it in v1 briefs? Proposal: keep it as the team's plan, renamed How your team measures it.
• v3 PRD
  v3 price. Proposal: the $149 or ₹9,999 working number in the pricing proposal [src: brew working tree docs/pricing.md:20].
• v4 PRD
  The whole document waits on you: Ads inside rules and a budget, undo, the rule language
• v4 PRD
  Which ad platform first? Proposal: Meta, since the connector and App Review work already exist.
• v4 PRD
  Default limits. Proposal: no defaults; the founder types the cap and step limit before any rule saves.
• v4 PRD
  What undo means for money spent. Proposal: undo restores settings only, and the log states the spend that already happened.
• v4 PRD
  Who may change the rules? Proposal: admins only, members may pause.
• v4 PRD
  Autopilot as a percent of spend or flat bands? Proposal: keep 3 percent with the $149 minimum as the working number until partners run it.
• v4 PRD
  Can an agency set rules per client? Proposal: yes, once the brand switcher exists, and the brand's admin can lock them.
• Roadmap with dates
  The whole document waits on you: What ships when, by version and quarter
• Roadmap with dates
  When does the design partnership end? Proposal: six weeks from the first partner's production onboarding on 26 Oct, then decide.
• Roadmap with dates
  Dates for steps 2 to 6. Proposal: one week per step after the brand direction is picked.
• Roadmap with dates
  When does pricing go live? Proposal: at the end of the partnership, with partners first.
• Roadmap with dates
  Is there a public launch day? Proposal: no single day; open reads in small batches as partners say it works.
• Roadmap with dates
  v2 trigger. Proposal: start v2 when the Meta review passes and two partners hand off weekly for three weeks.
• Personas
  Who is the first persona we design for? Proposal: the founder who is the marketing team.
• Personas
  Is the approver a role or just a member? Proposal: a member who lands on Need your yes, no new role chip.
• Personas
  Do we court agencies now? Proposal: not until the brand switcher; invite them per brand meanwhile.
• Personas
  Large brands like Puma India. Proposal: keep as a partner and learn, but design for small teams.
• Jobs to be done
  Is the main job right? Proposal: test it on the five partners' next calls before the redesign copy is written.
• Jobs to be done
  Do founders hire us for the Plays documents or for the feed and drafts? No usage tracking exists [src: docs/handover/product/07-open-questions-and-next-steps.md:35]. Proposal: count opens of Plays PDFs against jams for four weeks.
• Journeys
  Reveal before or after sign in? Today sign in comes first. Proposal: keep sign in first while reads are capped.
• Journeys
  Can an approver answer from email? Proposal: yes, with Yes and Not now links that open a confirm page.
• Journeys
  Can the person who ships mark a brief delivered? Proposal: no; only the brand's team and our team move status.
• Onboarding
  Which three questions? Proposal: the brief's three, and keep lead products as an optional Settings row.
• Onboarding
  Does an invited teammate get the questions? Proposal: only Which of these do you handle.
• Onboarding
  Can an agency start a read for a client? Proposal: not until the brand invites them.
• Onboarding
  The Shopify home says the first read takes about half an hour. Proposal: time it inside Shopify and use the same words as the web.
• Product rules
  AGENTS.md lets the handbook name the company behind our team, WRITING.md forbids it. Proposal: follow WRITING.md and say our team, and fix AGENTS.md.
• Product rules
  Is the 25 Oct Pro sheet exception still right? Proposal: drop it at the redesign.
• Design partner program
  The whole document waits on you: Who, what they get, what we ask, cadence, graduating
• Design partner program
  What do partners get? Proposal: free use through the partnership, our team on call, and a $120 or ₹10,000 wallet each.
• Design partner program
  What do we ask of them? Proposal: a 15 minute call each week, honest notes and at least one hand-off a week.
• Design partner program
  How long is the partnership? Proposal: six weeks from 26 Oct, then a keep going call.
• Design partner program
  How do they graduate? Proposal: to $79 or ₹4,999 a month for life, as in the pricing proposal.
• Design partner program
  Is Gully Labs still in? Proposal: no, the five are the list.
• Design partner program
  How many partners? Proposal: stay at five until each hands off weekly.
• Partner feedback log
  Partners' own words are missing. Proposal: a 15 minute weekly call per partner, notes logged here.
• Partner feedback log
  Where do Sai's and Arvind's notes go? Proposal: here, with their name in Who.
• Product metrics
  The whole document waits on you: Activation, first win, weekly use, hand-offs, defined
• Product metrics
  Is the north star right? Proposal: yes, weekly hand-off stores.
• Product metrics
  Who builds the decision log, and when? Proposal: engineering, as v1 groundwork before the redesign.
• Product metrics
  Targets. Proposal: none until four weeks of partner data.
• Product metrics
  Analytics tool. Proposal: our own event table in Postgres, no third party trackers.

Design 35

• Design principles
  Does truth outrank every brand as the top rule, or the other way round? Proposal: keep truth first, since a wrong number on one brand is worse than a missing feature on every brand.
• Design principles
  The live product is dark by default. Should the redesign keep dark as the default or follow the device? Proposal: follow the device, with dark as the fallback, because both themes ship for every direction.
• Design principles
  The glass tiers and the per brand wash are still in today's tokens. Remove both in the restyle? Proposal: yes, remove the glass tiers in v1 and keep the brand's photos as the only brand colour.
• Design principles
  Should the principles page carry a short example screen for each rule? Proposal: yes, one frame from the screen kit per rule, once a direction is picked.
• Components
  Fetch has extra poses for heart, drop and handoff that the kit never calls. Use them? Proposal: Use the heart pose on the first heart celebration at 120 px and drop on the Monday greeting; leave handoff for marketing.
• Components
  The felt plate is a book treatment. Should it sit behind Fetch in the product too? Proposal: No, keep the product calm; the plate stays in books, the site and marketing.
• Components
  Small buttons, chips, tools and segments are 34 to 36 px, under the 44 px phone target. Grow them? Proposal: Keep the look and give each a 44 px hit area.
• Components
  The five area colours stay as small dots beside the area name. Keep them in Fetch's skin? Proposal: Yes, as dots only, never as text or fills.
• Components
  The kit's hand off line for our team still names the old company. Rewrite? Proposal: Yes: "Our team has your brief now."
• Patterns
  Should Undo stay reachable after the toast goes, for example from the item's menu? Proposal: no, three seconds plus Later in See everything is enough for v1.
• Patterns
  Do the keys show a small hint the first time on a laptop? Proposal: yes, one line under the first pick, once.
• Patterns
  The hand off sheet lists four choices but the kit words describe three. Is Me a choice in v1? Proposal: yes, Me ships Copy and open, which already exists.
• Patterns
  Should Slack keep its make button until the Next change? Proposal: remove it in v1 so every paid make starts at the one confirm sheet.
• States
  Empty and error slots are 96 px, under the 120 px switch, so Fetch's empty and error poses never show in those slots. Raise those slots to 120 px or accept the small mark? Proposal: accept the small mark in the product and keep the full poses for marketing.
• States
  The kit has no words for See everything filtered to nothing. Proposal: "Nothing in this area this week." with Show all areas.
• States
  Should the first heart celebration use Fetch's full celebrate pose rather than 40 px in the toast? Proposal: yes, once, at 120 px, then the toast for every later heart.
• Content design
  Our team's future name is open, and the kit strings still name the old company. What is the plain word for our team on screens? Proposal: "our team" until the name is picked, then one find and replace.
• Content design
  Should card generation be shortened to the 8 and 12 word budgets, which is a core change? Proposal: ship the lint first, measure for two weeks, then decide.
• Content design
  Today's feed end says "You're done for today" with Find more. Tomorrow says caught up for the week. Is the week the unit for every brand? Proposal: yes, five picks a week, See everything for the rest.
• Content design
  Areas: are Social posts, Ads, Website, AI search, Email and WhatsApp the final five names? Proposal: accept them as written.
• Accessibility
  The whole document waits on you: Contrast, tap targets, motion, screen readers, checks
• Accessibility
  The 9.2 to 1 figure from the docs site audit has no file behind it, and recomputing the docs tokens gives 4.8 to 1 for muted text on Seeti light. Which elements did the audit measure? Proposal: rerun it, publish the script beside the docs and record ink and muted separately.
• Accessibility
  Four directions' hero colours are below 3 to 1 on the light page. Is an edge or shape on hero indicators acceptable? Proposal: yes, add a 1 px ink edge to the unseen ring and the on switch in light themes.
• Accessibility
  Should small buttons and chips grow to 44 px tall, or keep their look with a larger hit area? Proposal: a larger hit area, so rows stay quiet.
• Accessibility
  Who runs the first screen reader pass and when? Proposal: Sai, on the chosen direction, before the restyle ships.
• Character usage
  Fetch has extra poses for heart, drop and handoff that the kit never calls. Use them? Proposal: Use the heart pose on the first heart celebration at 120 px and drop on the Monday greeting; leave handoff for marketing.
• Character usage
  The felt plate is a book treatment. Should it sit behind Fetch in the product too? Proposal: No, keep the product calm; the plate stays in books, the site and marketing.
• Character usage
  The model puts the first heart in a 120 px celebration; the kit draws it at 40 px inside the toast. Which? Proposal: The first heart at 120 px once, later hearts in the toast.
• Character usage
  Should the character layer start on or off for a new brand? Proposal: On, with sound off, since the slots never block a control.
• Web and phone layouts
  Tablet has no kit frame. Draw one at 820 by 1180 before tablet ships? Proposal: yes, in v1, since the rules exist.
• Web and phone layouts
  Does the Shopify admin frame need a slimmer top bar? Proposal: test it against Shopify's App Bridge rules after the app review clears.
• Web and phone layouts
  Is 1024 the right switch to the rail, given today's most common break is 700? Proposal: keep 1024 from the model, and use two columns of cards between 600 and 1024.
• Design QA checklist
  Should 40 checks be one gate, or split into a fast gate for small changes and the full list for new screens? Proposal: a fast gate of checks 1, 2, 6, 9, 12, 20, 25, 36, 38 and 40 for small changes, the full list for new screens.
• Design QA checklist
  Where does the QA sheet live? Proposal: one row per screen and check in the handbook, linked from Every screen.
• Design QA checklist
  Is a headless audit worth building for checks 14, 36 and 38? Proposal: yes, reuse the loader audit approach with puppeteer.

Brand 39

• Naming
  The whole document waits on you: The rename: criteria, shortlist, trademark and domain checks
• Naming
  When do we name it? Proposal: right after the character pick, before any screen is reskinned, so the slot is filled once.
• Naming
  Which tests must a name pass? Proposal: accept the tests above as they are: one or two syllables, easy in every accent, says the job or sits well beside the character, not a known app, a free domain, a clean India trademark search.
• Naming
  Does the character name ever become the product name? Proposal: no. Every brand book already says the character is never the product, and a character can be swapped later without renaming the product.
• Naming
  Who does the trademark check? Proposal: a quick search by our team for the shortlist, then one paid opinion from a trademark lawyer in India on the final two, before any domain is bought.
• Naming
  Does the product move off joinsquareup.com? Proposal: yes, to a domain on the new name once it clears; keep feed.joinsquareup.com redirecting so design partners' links keep working.
• Naming
  Which character alternates stay live? Proposal: keep the two alternates only for the picked character, as a fallback if its name fails clearance, and archive the rest with their folders.
• The five directions
  Should Fetch get a critic round before the pick, as three others did? Proposal: yes, one fresh critic scores Fetch and Buzz against the bar and Octalysis, so all five are compared on the same footing.
• The five directions
  Is a happy dog sharp enough, after a generic friendly dog was turned down? Proposal: decide on the bad day and ten second tests with five founders, not on the drawing alone.
• The five directions
  Who are the deciders and by when? Proposal: Param, Sai and Arvind answer the Monday question alone this week, and Param picks after the founder tests.
• Fetch character bible
  Is Fetch sharp enough after a generic friendly dog was turned down before? Proposal: run the ten second and bad day tests with five founders and a critic round against the bar before the pick.
• Fetch character bible
  Is Fetch the final name, given it cannot be owned as a trademark alone? Proposal: keep Fetch as the character name, since it is never the product name, and hold Scruff as the alternate.
• Fetch character bible
  Who redraws the final art? Proposal: brief one illustrator from the rig and this bible after the pick, keeping the rig as the source for every product frame.
• Fetch character bible
  Which motion values win where the books disagree? Proposal: the system file values, 1 s celebrate and 2.4 s error, and fix the motion page to match.
• Buzz character bible
  Is Buzz the final name? Proposal: yes, Buzz, with Waggle kept as the fallback if a clash turns up.
• Buzz character bible
  Who redraws the final art if Buzz is picked? Proposal: one illustrator redraws from the rig and model sheet, keeping the two tells, and the rig is updated to match.
• Buzz character bible
  Which wait timing is right, 3.2 or 3.4 seconds? Proposal: 3.2 seconds, the value the screen kit already runs.
• Buzz character bible
  Two kit skin lines, handedOffMe and workNewIntro, use the word the Plays rule bans. Proposal: change them to Over to you and These changed.
• Scoop character bible
  Is gossip the right frame for the product? Proposal: keep it, because the read is our real edge and the public pages line turns the trust rule into a catchphrase, and test it with the five founder tests before deciding.
• Scoop character bible
  Is Scoop the final name, over Tidbit and Rummage? Proposal: keep Scoop, since the character and the weekly picks share one word.
• Scoop character bible
  Who redraws the final art? Proposal: brief one human illustrator with this bible, the master and the turnaround, and keep the generated set as reference only.
• Scoop character bible
  The redrawn hero and icon have not been scored. Proposal: run one fresh critic round on the new art before the pick.
• Ocho character bible
  Is Ocho the final character name? Proposal: keep Ocho, since it says the job in two syllables and works in every accent, and keep Inka as the fallback if a trademark clash appears.
• Ocho character bible
  Who redraws the art, and when? Proposal: if Ocho is picked, brief one illustrator to redraw only the master, the small mark and the ten poses from the rig before any marketing goes out.
• Ocho character bible
  Does the frantic expression stay out of the product entirely? Proposal: yes, frantic lives in marketing and on the stage only, as the books already say.
• Seeti character bible
  Is Seeti the final character name? Proposal: keep Seeti, with Whistle as the fallback if a trademark clash turns up.
• Seeti character bible
  Should Seeti stay it, or become she, since the voice is a woman? Proposal: it in every line we write, and let fans say she.
• Seeti character bible
  Who redraws the master if Seeti wins, and with what budget? Proposal: one illustrator for the master, the 24 pixel mark and the turnaround, then rebuild every frame from the rig.
• Seeti character bible
  Is boss, once a week, warm or too familiar outside India? Proposal: keep it once a week and test it in the bad day test.
• Visual identity
  What is the clear space around the lockup? Proposal: one ear width on every side, matching how the other directions set it by a body part.
• Visual identity
  Should the product light theme use the kraft book colours or the neutral kit light set? Proposal: keep the neutral kit set in the product and kraft for books and marketing, as the system file says.
• Visual identity
  When the name is picked, which face sets the wordmark? Proposal: Grandstander at 700, so the name and the headings share one voice.
• Sound and motion
  Should sound stay on by default as it is today, or start off as the Fetch books say? Proposal: off by default, on with one switch in Settings.
• Sound and motion
  Do Fetch's cues replace today's heart, ready and error cues, or play beside them? Proposal: replace, so one tap is still one sound.
• Sound and motion
  Is the voice used in the product at all? Proposal: only the paste line during onboarding, once, and never again in the product.
• The Brand Book we give brands
  Does Fetch sign the public Brand Book at all, or only the in product view? Proposal: sign both with the 32 px mark and one line, since the book is a gift from us.
• The Brand Book we give brands
  What do we call it to brands: Brand Book, or Fetch's word, the scent? Proposal: Brand Book on every control and in every file name, the scent only in Fetch's lines.
• Asset library
  Who draws the final art and the sticker? Proposal: one illustrator after the pick, working from the rig, with a sticker sheet in the same brief.
• Asset library
  Where do brand assets live once picked: brand-directions/dog or brew/public? Proposal: the rig stays in brand-directions, and only the built SVGs and mp3 files are copied into brew/public.

Engineering 39

• Architecture
  When do we run more than one worker? Proposal: stay on one worker until a build queue of more than 3 brands waits over 30 minutes, then build per-brand leases.
• Architecture
  When does staging stop mirroring to the bucket? Proposal: turn OBJECTS_MIRROR and OBJECTS_READTHROUGH off on 25 Nov 2026 if no rollback to the bucket happened.
• Architecture
  Should the demo move to Postgres? Proposal: yes, in v2, on its own small database, so the demo tests the same storage as production.
• Codebase map
  Should the big four files be split before v2? Proposal: split feed-service.js first (feed, jam routes, Slack in), one area per week, each behind the existing tests.
• Codebase map
  Should jam2/ be renamed now that the v2 screen is gone? Proposal: rename to studio/ in v2 with the v2-only routes removed in the same change.
• Data model
  Which domain tables come next? Proposal: connections and secrets first (they carry the protected-connection rule), then jams.
• Data model
  How long do we keep _deleted/ archives and raw/ fetches? Proposal: 30 days for _deleted/, 90 days for raw/, by bucket lifecycle rules.
• Environments
  Is Arata's home production or staging? The settings say production; two comments say staging. Proposal: production, and fix the two comments.
• Environments
  When does production get its own domain? Proposal: keep feed.joinsquareup.com until the brand name is picked, then move with a redirect.
• Environments
  Should Sloosh finals run on production? Proposal: no until the Sloosh workspace has paid credits and Param approves the first quote there.
• The read
  What does a read cost on Gemini today? Proposal: measure the next Cozy re-read and record it here.
• The read
  Should a signed-out visitor still start a read? Today non-official emails cannot start a new read. Proposal: keep it so; the demo is the place for prospects. [src: memory brew-access-rules]
• Plays engine
  Which model writes Plays at launch? Proposal: Opus when Anthropic credit exists, since the 20 Oct comparison favoured it; run the fair 5-Play test first (about $1).
• Plays engine
  Should an unverified Play show to the brand? Proposal: yes, with "N claims to review" as today, never hidden.
• Cards and the feed
  Are the ranking weights right? Proposal: build fixtures from the five partners and freeze the top 20 per brand as a test, then tune.
• Cards and the feed
  Should Storefront come back for v2? Proposal: yes for Shopify brands with a connected store only.
• The jam and Studio
  When do the v2-only routes go? Proposal: check the access log for /p/ and /brief/ hits over 14 days; remove them if there are none.
• The jam and Studio
  Does Set-up v2 go to production? Proposal: yes after one partner session on staging with no layout findings.
• Briefs and hand-off
  Should shared links expire? Proposal: 30 days by default, with revoke on the jam.
• Briefs and hand-off
  Does the recipient "our team" stay now that Brew is kept independent? Proposal: keep it as is, add nothing new tied to it. [src: memory brew-qa-round2-2026-10-24]
• Connectors
  When should Marketplaces become real? Proposal: after Meta App Review, starting with Amazon seller data for India.
• Connectors
  Who besides Arata goes on the protected list? Proposal: every design partner once their first account connects.
• Learnings and memory
  Should brand memory and Brand Book rules merge? Proposal: not in v1; in v2 make memory lines that are rules become Brand Book rules after a team confirm.
• Learnings and memory
  Should a member's preferences ever reach the brand's memory? Proposal: no, they stay personal.
• Testing and QA
  Should a red CI run block a deploy? Proposal: yes, deploy.sh checks the commit's Actions status first and stops on red.
• Testing and QA
  Who owns QA rounds once Sai and Arvind work in the code? Proposal: whoever ships a phase runs its QV round; Param signs off the three final rounds.
• Release and promote
  Who may run a promote? Proposal: Param, or a session Param asks in chat, one at a time.
• Release and promote
  Is there a release freeze during a partner session? Proposal: no promotes during a booked partner call; staging stays free.
• Release and promote
  What does production run now? Memory says 0285592 on 25 Oct. Proposal: print the serving sha on /api/config so this page can read it.
• Parallel sessions
  Should sessions move to branches and pull requests now? Proposal: yes for any change over 200 lines; small fixes stay on main.
• Parallel sessions
  Should long paid runs block deploys? Proposal: no; keep resumable runs and never re-add a deploy wait.
• Decision records
  Which model writes Plays at launch? Proposal: Opus with Anthropic credit, after a 5-Play blind test against Pro (about $1).
• Decision records
  Should DR-17 also cover Sloosh finals on production? Proposal: yes, keep finals on Gemini on production until Param says otherwise.
• Tech debt
  Which High item first, now that CI runs? Proposal: D-1 this week, then D-3.
• Tech debt
  Can D-1 be fixed without breaking the other services? Proposal: list the 15 services and give each its own account first, then remove the default grants.
• Engineering rules
  Should these rules be enforced in CI where possible? Proposal: yes; start with the Dummy-marker check and a ban on brand slugs in code outside tests and data.
• Performance and reliability
  What uptime do we promise design partners? Proposal: 99.5% a month for the feed, measured by the /healthz/ready check, stated in the partner terms as a target, not a guarantee.
• Performance and reliability
  What page speed is good enough? Proposal: feed under 500 ms at p95 and brand page first paint under 2 s on a phone, both checked after each promote.
• Performance and reliability
  Should we load test again before more brands join? Proposal: yes, at 3 times the 25 Oct rate (about 90 requests a second) before the next onboarding wave.

AI and models 31

• Model policy
  Back to Claude for Plays? Proposal: top up Anthropic, run the fair bake-off from the Evals document on 5 Arata Plays, about $5 to $10, and switch Plays back to Opus medium only if Opus wins most blind pairs.
• Model policy
  Is Claude on Vertex covered by the Google credits? It has never been used [src: memory brew-gemini-provider]. Proposal: one small paid probe on Vertex with your yes, then decide whether Claude runs on the credits.
• Model policy
  Brief and ideas on Flash. They asked for Opus high and get Flash today. Proposal: move the brief to Pro alongside Plays, since every Play reads it.
• Model policy
  Jam spend has no dollar meter. Jam calls count only as daily units [src: brew/server.js:146-171]. Proposal: meter them through ai.js so cost per brand includes the jam.
• Prompts registry
  Should the registry become code? Proposal: yes, a data/prompts.json read by ai.js and server.js, in v2, with a test that every call site is listed.
• Prompts registry
  Which tags need a stronger tier first? Proposal: brief and ideas:write move to Pro with the Plays, because they asked for Opus and every Play reads the brief.
• Guards
  Count the guards. Nothing records how often each guard fires. Proposal: one daily line per brand in the ops page with flags, mends, redos and holds.
• Guards
  Judge model. The picture checks ask for Haiku and run on Flash today. Proposal: keep Flash until the eval set shows judge agreement with a person below 90%, then move the judge to Claude.
• Guards
  Video spelling remake. Proposal: keep it manual in v1, and in v2 offer one priced remake.
• Evals
  The whole document waits on you: What we measure on outputs, test sets, cadence
• Evals
  Accept this eval set? Proposal: accept as written, run the first Play bake-off on 5 Arata Plays across Opus, Pro and Flash before deciding the Plays model.
• Evals
  Who scores blind pairs? Proposal: Param and one of Sai or Arvind each score every pair; a tie between them counts as a tie.
• Evals
  Open models in the bake-off? Proposal: leave them out of the first run; add one when the gates need a cheaper classifier.
• Evals
  Use real partner packs? Proposal: yes, as frozen read-only copies that never leave our machines and never write back.
• Evals
  Eval budget. Proposal: $20 a month while building, from the Google credits for Gemini runs, with your yes before any Claude run.
• Image and video makers
  Higgsfield keys. Proposal: add HF_API_KEY_ID and HF_API_KEY_SECRET to Secret Manager so Studio runs on its first choice, with the existing per-make price and yes.
• Image and video makers
  What is a Sloosh credit in dollars? No source gives it. Proposal: write the plan price into Finance and show finals in dollars to the team.
• Image and video makers
  Sloosh or Higgsfield for finals? Both make finals-grade work. Proposal: run the partner bench on 10 briefs and keep the cheaper one that passes the eval bar.
• Image and video makers
  Higgsfield Enterprise for the CLI-only tools? Needs a word with Higgsfield sales before Brew wires Marketing Studio video or audio. Waits on Param [src: memory brew-higgsfield-wiring].
• Outside data rules
  The skill is not in the repo. It lives in the parent folder, not on brew origin/main. Proposal: move it into brew/.claude/skills so every Brew session loads it.
• Outside data rules
  One cap per brand. treg is capped per day for all brands, Apify per brand per month. Proposal: add a per-brand monthly treg cap of $15 beside Apify's $10, both shown on the watch page.
• Outside data rules
  The Apify token was pasted in chat on 21 Oct. Proposal: rotate it [src: memory brew-viral-remakes].
• Cost per brand
  Measure one real month. No brand has a measured monthly total. Proposal: add AI dollars per brand to the spend file, then report October and November for Arata, Nutrova and Cozy.
• Cost per brand
  What is a Sloosh credit in dollars? Proposal: write the Pro plan price into Finance and convert finals on the ops page.
• Cost per brand
  Read cost. The first read has no single figure. Proposal: tag every read call reader-<slug> and sum it on the next new brand.
• Cost per brand
  Cap for Plays on Opus. Proposal: if Plays go back to Opus, set BUILD_CAP_USD to $90 for that deploy.
• Safety and refusals
  Product name in messages. Some messages say "Brew", and some name Higgsfield, while product rules say no product name and no vendor names on founder screens [src: brew/drafts.js:183] [src: brew/jam2/studio.js:301] [src: AGENTS.md rule 6]. Proposal: one copy pass over every refusal and budget message.
• Safety and refusals
  Refusal log. Refusals are kept only in memory on one instance. Proposal: store them per brand and month beside the spend file.
• AI search visibility
  Claude in the check. Claude sits out while Brew runs on Gemini, so the check shows three tools. Proposal: ask Claude through the Anthropic key for this one check even while the rest runs on Gemini, about 8 short calls a week a brand.
• AI search visibility
  Weekly check across many brands. About $0.50 a brand a week. Proposal: keep it weekly for design partners and monthly for brands that only had a read.
• AI search visibility
  Google AI overviews. Not asked today. Proposal: probe one treg endpoint for AI overviews on one brand with your yes, then add it if it answers reliably.

Data and analytics 20

• Metric definitions
  The whole document waits on you: Every number the product and company use, defined once
• Metric definitions
  Accept the proposed company definitions? Proposal: yes as written, with the north star counting any hand-off, to the brand's team or to an outside maker.
• Metric definitions
  Does staff activity ever count? Proposal: never; our team's opens and makes are counted apart.
• Metric definitions
  Cost per active brand: include Google credits? Proposal: yes, count credits as money, so the number holds once credits end in 2028.
• Measurement engine
  Which grades teach the feed? Proposal: A and B only; C and D show to the team. [src: handbook content/company-and-strategy/open-questions.md:37]
• Measurement engine
  Keep How we measure it in v1 briefs? Proposal: keep it, renamed How your team measures it. [src: handbook content/company-and-strategy/open-questions.md:40]
• Measurement engine
  Approve Other proposals without a route? Proposal: leave them stored until v3 brings the engine back.
• Tracking plan
  The whole document waits on you: Events, properties, where they fire
• Tracking plan
  Approve the event list? Proposal: yes, the v1 rows only, built as one change by engineering before the redesign.
• Tracking plan
  Our own table or an analytics vendor? Proposal: our own table in Postgres, so no new subprocessor and no cookie banner.
• Tracking plan
  Track the future public site? Proposal: the marketing site may use the Setup Analytics pattern; the product itself never does, and the privacy page says which is which.
• Dashboards
  The whole document waits on you: What each shows and who reads it
• Dashboards
  Build the weekly views in the product or in a tool? Proposal: a small staff page in the product, so data never leaves our database.
• Dashboards
  Who owns the Monday read? Proposal: Sai prepares, Param decides.
• Dashboards
  Show partners their own view? Proposal: not in v1; in v3 the brand gets its results view.
• Data quality
  Run a nightly data check? Proposal: yes, one report to hello@ listing each partner's unbacked numbers and stale connections.
• Data quality
  How long may a Play stay flagged? Proposal: 7 days, then it leaves the feed until it is refreshed.
• Experiments
  Holdouts through the brand's own tools? Proposal: yes, Klaviyo and the brand's WhatsApp tool first, never our own sending. [src: handbook content/company-and-strategy/open-questions.md:39]
• Experiments
  Test our own product before v3? Proposal: no; with five partners, compare weeks before and after a change and ask them.
• Experiments
  Default holdout share? Proposal: 10%, shown to the brand before launch.

Ops and infrastructure 33

• Runbook index
  Who may run production runbooks? Today the scripts run with whoever's gcloud sign-in is on the Mac. Proposal: only Param runs promote.sh and the restore runbooks until a second person has done one restore test with him.
• Runbook index
  Should the bucket scripts be deleted now? tools/reset-brand.sh and tools/flush-live.sh already stop on STORAGE=pg. Proposal: give tools/reset-jams.sh the same first lines, and delete all three once production has run on Postgres for a month.
• Deploy and promote
  How often should production be promoted? No rhythm is written down; Next carries one pending promote of the 26 Oct jam fixes. Proposal: promote at most once a day, after Param has walked Arata and one other partner on staging.
• Deploy and promote
  Should staging smoke checks fail the deploy? Today they only print codes. Proposal: yes, fail on any non-200 from / or /api/config.
• Restore
  Who runs the monthly restore test, and on which day? Proposal: Param, on the first Monday of each month, starting 2 Nov 2026, with Claude doing the steps and Param checking the counts.
• Restore
  Is staging data worth backing up? The partners now live on production. Proposal: treat staging as rebuildable, keep only the bucket versioning, and say so here.
• Restore
  Is 7 days of point-in-time recovery enough? Proposal: keep 7 days during the design partnership and revisit when the first partner pays.
• Wipe data and reset jams
  May a brand reset run on production? The code allows it, with no archive. Proposal: no; production runs only db/delete-brand.mjs for a data request.
• Wipe data and reset jams
  What clears the cache and made media on Postgres? Proposal: add it to db/reset-brand.mjs behind a --with-cache flag, so "wipe data" is one job.
• DNS and domains
  Who owns the Cloudflare account and the joinsquareup.com zone? No note names it for this zone. Proposal: Param owns it, with Sai as a second admin so a lost login never locks the domain.
• DNS and domains
  When do we move production off feed.joinsquareup.com? Proposal: when the product name is picked, in one planned change with every callback URL checked from the list above.
• Secrets and keys
  Who rotates which secret? No source names anyone. Proposal: Param owns every secret for now; Claude may add versions only when Param asks in chat; a review every quarter starting January 2027.
• Secrets and keys
  Should production have its own AI and data keys? Today staging, production and the demo share ANTHROPIC_API_KEY, OPENAI_API_KEY and TREG_TOKEN. Proposal: yes for treg and Anthropic before the first paying partner, so a staging leak or runaway test cannot spend production's balance.
• Secrets and keys
  Fix the default compute account's access? Proposal: yes, remove project-wide secret accessor and give each of the 15 services its own account, in one planned session.
• Monitoring and alerts
  Who reads hello@ and who answers out of hours? No source names anyone. Proposal: Param is on call every week, with Arvind as backup, until a second engineer joins; hello@ forwards to Param's phone; any alert gets a first look within 2 hours from 9 am to 10 pm IST.
• Monitoring and alerts
  Should staging have an uptime check? Proposal: yes, one check on /healthz/ready, emailing hello@ only.
• Monitoring and alerts
  Should the Sloosh check run on production? It is off there because Sloosh is not connected on production. Proposal: turn it on the day Sloosh is connected there.
• Incidents
  The whole document waits on you: Calling one, roles, the postmortem template
• Incidents
  Accept the proposed process above? Proposal: accept as written; Param is the lead for every Sev 1 until a second engineer joins.
• Incidents
  Where do we talk during an incident? Proposal: a Slack channel #incidents in the joinsquareup workspace, plus a phone call for Sev 1.
• Incidents
  What do we promise partners about incidents? Proposal: for Sev 1, a message within an hour and a short written postmortem within 2 working days; nothing about uptime until we measure it.
• Costs and quotas
  The whole document waits on you: Cloud, AI, data, credits, monthly spend
• Costs and quotas
  Who checks spend, and how often? Proposal: Param, on the 1st of each month, copying each vendor's last-month total into Numbers; Claude drafts the table from the billing pages he opens.
• Costs and quotas
  What is the monthly ceiling while we are pre-revenue? Proposal: the ceilings proposed in budget: a Google Cloud budget alert at $1,000 a month, its cloud $400 plus AI $600, and $220 for treg and Apify together, both emailing hello@.
• Costs and quotas
  Do the production caps fit five partners? The caps are the staging numbers copied. Proposal: keep them until the first month's spend is known, then set production's from real use.
• Costs and quotas
  Which vendors are we paying today, on which card? Not recorded. Proposal: Param lists them once in Finance, and this page links there.
• Protected connections
  Protect the other four partners too? Today only Arata is listed. Proposal: add each partner the day it connects its first real account, starting with Nutrova.
• Protected connections
  Should force exist at all on production? Proposal: keep it for Param only, log every use, and require a note saying why.
• Background jobs
  Are these cadences right for paying partners? They were set during building. Proposal: keep them for the design partnership and review after a month of spend data.
• Background jobs
  Should weekly work on production run at a fixed hour? Today each job's first run depends on when the worker started. Proposal: no change until a partner asks for a fixed time.
• Hosting map
  When does the handbook go live? The deploy script already keeps the pages off public hosting and refuses a release if squareup-handbook.web.app, firebaseapp.com or run.app answer a signed-out request, but nothing is uploaded until Cloudflare Access is on. Proposal: Param switches on Access for handbook.joinsquareup.com this week and sends the team domain and audience tag, then the first release runs with the gate, never with --skip-gate.
• Hosting map
  Who besides Param gets into the handbook? Proposal: Access allows Param, Sai and Arvind by email; the comments service's ALLOWED_EMAILS holds the same three.
• Hosting map
  Should we clean up the old buckets gs://squareup-feed and gs://squareup-feed-demo? Proposal: export a copy to cold storage, then delete them after Param confirms nothing reads them.

Security, privacy and legal 52

• Privacy policy
  Who reviews the final policy, and when? Proposal: a lawyer in India who knows the Digital Personal Data Protection Act, before the first brand outside the five partners signs up.
• Privacy policy
  Which laws apply to brands outside India? Cozy is a US store. Proposal: the lawyer confirms whether GDPR and US state laws need extra sections; until then, onboard only Indian brands to production.
• Privacy policy
  Do we need a cookie banner? Proposal: no, while the only cookies are the two sign in cookies and there are no trackers; the lawyer confirms.
• Privacy policy
  Does the policy name the product? The writing rules say no product name on customer pages. Proposal: the policy names the legal entity and says "the service" until the brand name is picked.
• Privacy policy
  Should we keep the Anthropic key on production when nothing uses it? Proposal: remove it, so the subprocessor list stays short and true.
• Terms of service
  Do people accept the terms, or only see them? Nothing records acceptance today. Proposal: one line under the sign in button, "By signing in you accept the terms", plus the date stored on the brand when its first admin claims it.
• Terms of service
  Which city's courts? Proposal: the city on the LLP's registration; Param confirms.
• Terms of service
  Liability cap? Proposal: fees paid in the last 12 months, and a fixed small amount during the free preview; the lawyer sets the number.
• Terms of service
  Who reviews, and when? Proposal: the same lawyer as the privacy policy, in the same review, before the first brand outside the five partners.
• Data processing agreement
  The whole document waits on you: For partner brands
• Data processing agreement
  Do we offer a DPA now, or only when a partner asks? Proposal: now, with the partner agreement, because PUMA India is likely to ask.
• Data processing agreement
  Breach notice time? Proposal: 72 hours from confirming, which is a common standard; the lawyer checks Indian law.
• Data processing agreement
  Backups after deletion? Point in time recovery keeps 7 days and daily backups keep 14. Proposal: say backups roll off within 14 days and are never restored for a deleted brand.
• Data processing agreement
  Who signs for us? Proposal: Param, as a designated partner of the LLP; he confirms he is one.
• Subprocessors
  Which region does Vertex run Gemini in? Proposal: pin it to asia-south1 where the model exists there, and say so on the page.
• Subprocessors
  Is Higgsfield live on production? Proposal: Param confirms whether HF_API_KEY_SECRET exists; if not, drop the row until it does.
• Subprocessors
  Is the Cloudflare record proxied? Proposal: keep it DNS only, so Cloudflare sees no traffic, and confirm in the dashboard.
• Subprocessors
  Do we publish vendor names on a public page? Proposal: yes, a short list with purpose and country, the way most software companies do.
• Retention and deletion
  Brand data with no end date: do we set one? Proposal: 12 months without any sign in, then an email to admins, then deletion 30 days later.
• Retention and deletion
  How long do we keep the access log? Proposal: 12 months.
• Retention and deletion
  Do restored backups bring deleted brands back? Proposal: keep a list of deleted slugs and delete them again after any restore.
• Access rules
  Should every shopos.ai address see every brand? Proposal: no; a named list of staff addresses on production, and the two domains only on staging.
• Access rules
  Turn on STAFF_REQUIRE_HD=1 on production? Proposal: yes, so a personal Google account made with a staff address never counts as staff.
• Access rules
  Should guest links expire? Proposal: 30 days, with a Revoke button for brand admins.
• Platform compliance
  Meta verification has been in review 10 days. Chase it? Proposal: open a support case in Business Help if it is still in review on 30 Oct.
• Platform compliance
  Move the Shopify app's URLs to production before approval? Proposal: no, change nothing during review; move them the day it is approved.
• Platform compliance
  Point Meta's privacy URL at production now? Proposal: yes, feed.joinsquareup.com/privacy, since the page is the same and production is the brands' home.
• Claims policy
  Who checks the claim rules against Indian advertising law? Proposal: the same lawyer as the privacy policy reads this page and the category word lists once, before the first brand outside the partners.
• Claims policy
  Do creator drafts carry a disclosure by default? Proposal: yes, "#ad" on any draft written for a creator to post for the brand.
• Claims policy
  Log every flag? Proposal: yes, one row per flag with brand, kind and surface, so we can see what writers get wrong.
• Security practices
  Fix the default compute account now? It touches other company services. Proposal: yes, this week, one service at a time, each moved to its own account first.
• Security practices
  An outside penetration test? Proposal: one after v1 ships to the first 20 brands, not before.
• Security practices
  Rotate the keys pasted in chat on 28 Sep? The SquareUp tracking keys share this Google project. Proposal: rotate them now. [src: memory squareup-tracking-structure]
• Incident response
  The whole document waits on you: Breach steps, who to tell, timelines
• Incident response
  Who is on call? Proposal: Param in the day, Arvind as backup; one phone number each in the team channel.
• Incident response
  Do we promise brands 72 hours? Proposal: yes, in the DPA, after the lawyer confirms Indian law needs nothing shorter.
• Incident response
  A security@ address? Proposal: yes, forwarding to Param and hello@, named on the privacy page.
• Incident response
  Rotate the 28 Sep keys now? Proposal: yes, before any brand asks about security.
• Trademark and IP
  The whole document waits on you: The name, the characters, rights to generated art
• Trademark and IP
  Search before or after the pick? Proposal: a quick search on all five names now, so a name that cannot be protected leaves the audition early.
• Trademark and IP
  Who made the character art so far? Proposal: Param lists each source; anything made by a freelancer gets a signed assignment.
• Trademark and IP
  Keep the viral clip as a motion reference? Proposal: keep it on the partners only until the lawyer has reviewed it.
• Trademark and IP
  Which entity owns the marks? Proposal: Yellow Monk Technologies LLP.
• Company entity and filings
  The whole document waits on you: The legal entity, registrations, GST, annual filings, who holds what
• Company entity and filings
  Is Form 8 for the year to 31 March 2026 filed? It is usually due 30 October. Proposal: Param asks the CA today.
• Company entity and filings
  Who are the designated partners? Proposal: Param fills the table; the same people sign contracts and the DPA.
• Company entity and filings
  Is the LLP registered for GST? Proposal: ask the CA before any paid plan; register when the CA says the turnover or export of services needs it.
• Company entity and filings
  Is SquareUp a trading name of the LLP? The domain is joinsquareup.com. Proposal: say "Yellow Monk Technologies LLP, trading as SquareUp" on legal pages if the CA agrees.
• Contracts register
  The whole document waits on you: Every signed agreement: partners, vendors, contractors, NDAs, with renewal dates
• Contracts register
  Are there agreements this page does not know of? Proposal: Param lists any NDA, contractor or freelancer agreement, including anyone who made the character art.
• Contracts register
  Save copies of vendor terms? Proposal: yes, a PDF of each vendor's terms on the day this register is filled, for the five vendors with brand data first: Google, OpenAI, Sloosh, Resend and treg.
• Contracts register
  Where is the signed copies folder? Proposal: a shared drive folder named Contracts, owned by Param, read by Sai and Arvind.

Marketing and growth 71

• Go to market
  India first, or global from day one? Proposal: India first, because all five partners are Indian brands and the regional price is in rupees; copy stays in plain global English.
• Go to market
  Who are the first 100 brands, and how do we reach them? Proposal: a written list per six week cycle from our own networks, partner referrals and the Shopify listing, with 40 names in the first cycle, as the old plan suggested.
• Go to market
  When is the launch? Proposal: no date until every partner hands off at least once a week for four weeks in a row; then open reads in batches.
• Go to market
  Who on the team owns outreach and the same day reply? Proposal: Param owns both, with one named backup from Sai or Arvind.
• Go to market
  Is Gully Labs still a design partner? Proposal: no; the five on production are the list.
• Go to market
  Do Shopify installs skip the cap? Proposal: no; an install is a read ask like any other.
• Ideal customer
  Do large brands like Puma India count as a target, or a learning partner only? Proposal: learning partner only; keep the written fit for outreach.
• Ideal customer
  Do enquiry stores like Mana stay in scope? Proposal: yes for partners already in, no new enquiry stores in outreach until v1 ships.
• Ideal customer
  Is there a store size floor? Proposal: no floor; measure the review count below which Plays get worse across the five partners, then decide.
• Ideal customer
  India only for outreach? Proposal: yes for the first cycle, as in go to market.
• Audiences and hooks
  May we interview partners for their own words this cycle? Proposal: yes, five questions on the week one call, quoted only with written permission.
• Audiences and hooks
  Do the landing sample cards stay? Proposal: label them Sample now, and swap in one real partner finding once a partner agrees in writing.
• Audiences and hooks
  Which audience gets the first paid test, if any? Proposal: none until paid acquisition is decided; organic first.
• Website
  Which domain does the site live on? Proposal: pick it with the name; until then keep feed.joinsquareup.com.
• Website
  Are the home page and the ask page one page? Proposal: one page; the ask is the last block, and a full cap swaps the box for Leave your link.
• Website
  Do the landing's Sample cards stay, now that they are labelled? Proposal: keep them until a partner lets us show a real finding.
• Website
  When does the internal name leave the landing and the emails? Proposal: the day the name is picked, in one change.
• Website
  Do we show one finding before sign in? Proposal: yes, as the reveal in v1, so a visitor sees value before giving an email.
• Content engine
  Which store do our sample drafts come from? Proposal: build the sample candle store as a test brand, labelled Sample, so no partner shows without permission.
• Content engine
  Which format retires first if time is short? Proposal: keep the Monday and Friday formats always; drop the monthly ones first.
• Content engine
  Who makes the posts? Proposal: the hours in the no-budget playbook, with one owner per weekday.
• Social
  What is the handle? Proposal: the product name, picked with the domain in naming.
• Social
  Which founder accounts post on LinkedIn? Proposal: Param from week one; Sai and Arvind one post a month each.
• Social
  When does the first post go up? Proposal: the week the direction and name are picked, starting with post one.
• Social
  Do we post partner work? Proposal: only with written permission, never before week four of a partnership.
• Launch plan
  What is the launch date? Proposal: none until every partner hands off weekly for four weeks; then open reads in batches.
• Launch plan
  What do partners get on paper? Proposal: free use through the partnership and a loaded wallet, as in design partner program.
• Launch plan
  When does the six week calendar start? Proposal: the Monday after the direction and name are picked.
• Launch plan
  Is there a public launch day at all? Proposal: no single day; each version gets its own small opening.
• Launch plan
  Who sends the Friday founder note? Proposal: Param, every Friday from week four.
• Lifecycle
  Who sends the person written emails? Proposal: Param, from his own address, until a support inbox exists.
• Lifecycle
  Do the day 1, day 2 and day 14 emails start now, before the Monday drop ships? Proposal: no; start with the welcome, read ready and day 28 by hand.
• Lifecycle
  Does the Monday digest become the Monday drop? Proposal: yes, one Monday message, never two.
• Lifecycle
  When do pushes start? Proposal: with the installed app and the Monday drop, under Fetch's ceiling above.
• Lifecycle
  What is the paying message? Proposal: written when pricing is agreed with Sai; until then, cost is talked about in person.
• Word of mouth and referrals
  Does the brief page carry Fetch's sign off? Proposal: yes, 32 px at the bottom, and nothing else.
• Word of mouth and referrals
  Can a brand switch its public brand book off? Proposal: yes, from Settings, with no need to ask us.
• Word of mouth and referrals
  Do we ever reward referrals? Proposal: no rewards before launch; revisit once pricing exists, and never as a discount.
• Word of mouth and referrals
  Which founder communities first? Proposal: three Param already belongs to.
• Paid acquisition
  The whole document waits on you: When, where, budget, rules
• Paid acquisition
  Do we do paid acquisition at all before launch? Proposal: no; organic and warm intros only while reads are capped.
• Paid acquisition
  What are the gates for the first test? Proposal: pricing agreed with Sai, the name chosen, and every partner handing off weekly for four weeks.
• Paid acquisition
  Which channel first? Proposal: Meta, Instagram only, in India, because the founder audience scrolls there and the ad account exists.
• Paid acquisition
  What budget? Proposal: Param names one fixed cap for a two week test; nothing is spent past it.
• Paid acquisition
  Do we set up tracking now, before any spend? Proposal: yes; run Setup Analytics for the product's own pages this cycle, at no cost.
• Paid acquisition
  Does the Cozy campaign teach us anything here? Proposal: one rule only: never launch paid without a working purchase or ask event to count.
• Getting found ourselves
  Which domain will the product live on? Proposal: decide it with the name in naming; keep feed.joinsquareup.com as a redirect.
• Getting found ourselves
  Do we run our own AI answers check before we have a name? Proposal: no; wait for the name, then run it weekly.
• Getting found ourselves
  May the docs site be indexed? Proposal: yes, once it names the product; today it stays as it is.
• Getting found ourselves
  Who writes the answer pages? Proposal: Param drafts with Claude, Sai reviews for plain words.
• Campaign log
  The whole document waits on you: Every campaign, cost, result
• Campaign log
  What counts as a campaign? Proposal: anything with a start, an end and one ask: a paid test, a six week posting cycle, an outreach batch, an event or a launch.
• Campaign log
  Who keeps the log? Proposal: whoever owns the campaign fills its row; Param checks it every Friday.
• Campaign log
  Does the log live here or in a sheet? Proposal: here, so every campaign sits next to the plan that made it.
• Campaign log
  Do we log the first design partner outreach after the fact? Proposal: yes, one row per partner with how they came in, from what Param remembers, marked From memory.
• Growth metrics
  The whole document waits on you: Visits, reads, sign ups, activation, cost per brand
• Growth metrics
  Which three do we count first? Proposal: read asks, reads started and weekly hand off stores.
• Growth metrics
  Do we add analytics to the landing and the ask page? Proposal: yes, one GA4 property with four events: ask, sign in, read started, first hand off; no ad pixel until paid is decided.
• Growth metrics
  Do team hand offs count? Proposal: yes; log a quiet event when Share with my team copies a brief, so both hand offs count the same.
• Growth metrics
  Who counts on Friday? Proposal: Param, until a dashboard exists.
• Growth metrics
  When do targets come? Proposal: after four weeks of counts, Param sets one target for weekly hand off stores only.
• No-budget playbook
  Who owns which hour? Proposal: Param the Monday and Friday hours, Sai the Wednesday hour, each of the three one LinkedIn post a month.
• No-budget playbook
  Which store do we run our own product on for sample drafts? Proposal: none of the five partners; build the sample candle store as a test brand, labelled Sample, so drafts never show a real partner without permission.
• No-budget playbook
  May we spend on a studio make for a post? Proposal: no; drafts only, and Param says yes to any exception in writing.
• No-budget playbook
  When does posting start? Proposal: the day the direction and the name are picked, never before, so the grid starts in one voice.
• Events and webinars
  The whole document waits on you: Store owner meetups, webinars, demos at events, what each is for
• Events and webinars
  Do we do events before launch at all? Proposal: one small founder meetup a month at most, only where founders of our ideal customer gather.
• Events and webinars
  Which communities first? Proposal: Param lists three founder groups he already belongs to; we start there.
• Events and webinars
  Webinars, yes or no? Proposal: not until one partner agrees to show their week on screen.
• Events and webinars
  Who speaks? Proposal: Param, with a speaker bio written from his real background, never a sample bio.
• Events and webinars
  Budget? Proposal: zero; free events and online only, until paid acquisition is decided.

Sales and partnerships 42

• Arata
  Who at Arata do we work with? No contact name is on record. Proposal: add the main contact, their role and how they prefer to talk to this page.
• Arata
  What is the next step with Arata? Proposal: a 30 minute call this week to walk the Plays they rated highest, then agree a design partner agreement from Contracts.
• Arata
  Does Arata stay on the demo site? It agreed on 15 Oct that its team notes stay public there, but it now has real accounts connected. Proposal: keep only its public pack on the demo and ask Arata to confirm in writing.
• Arata
  Is our team's deck for Arata a paid engagement? The sources do not say. Proposal: record whether Arata pays our team separately, so the Brew partnership stays independent.
• Nutrova
  Did the 26 Oct onboarding happen, and with whom? Proposal: add the date, the people and what they said to this page and the feedback log.
• Nutrova
  Which account does Nutrova connect first? Proposal: Shopify through its own app, since it needs no platform review and feeds the most Plays.
• Nutrova
  Who signs off health claims? Proposal: ask Nutrova for the studies behind its claims and load them into its Brand Book before any Studio make is shared.
• Mana
  Who is our contact at Mana, and how did they come to us? Proposal: add the contact and the date of the first conversation here.
• Mana
  Is an enquiry brand a fit for the first paid plan? Proposal: keep Mana as a design partner and decide fit after it has used the feed for four weeks.
• Mana
  Which account does Mana connect first? Proposal: GA4, by adding the service account as a viewer.
• Puma India
  Who at PUMA India is the contact, and who approves the use of PUMA's name on our side? Proposal: record the contact here and ask in writing before PUMA appears in any demo, deck or case study.
• Puma India
  Should a brand this big be a design partner? Proposal: keep it as the big brand test case, and never quote it as a customer until it says yes in writing.
• Puma India
  Do we fix the stray words picture first? Proposal: yes, trace the maker before PUMA's next session.
• Scrubroom
  Who is our contact at Scrubroom? Proposal: add the contact and the date of the first conversation here.
• Scrubroom
  Does Scrubroom want Plays about its own labels or the brands it resells? Proposal: ask on the first call; until then, keep both.
• Scrubroom
  Is it on Shopify? The sources do not say outright; the reader got its full catalogue from a products.json list, which Shopify stores serve [src: brew/reader/index.js:244]. Proposal: confirm on the first call, then connect its own app.
• Pipeline
  The whole document waits on you: Every prospect, stage, next step
• Pipeline
  Who else are we talking to? Proposal: Param lists every brand in conversation, with the contact and the next step, and we add them here this week.
• Pipeline
  Is Gully Labs still a design partner? Proposal: mark it Paused unless Param says it is active.
• Pipeline
  Do these seven stages fit? Proposal: accept them as written and revisit after the first two signed agreements.
• Pipeline
  Who owns each prospect? Proposal: Param owns all of them until a second person takes one.
• Pipeline
  Where does the pipeline live? Proposal: here in the handbook only, no separate CRM until there are more than twenty rows.
• Onboarding playbook
  Is the first hand off the activation we count? Proposal: yes, first hand off within 7 days of the first sign in.
• Onboarding playbook
  Does every partner onboard on a call, or alone? Proposal: on a call for the five design partners; alone, with the first win bar, for everyone after.
• Onboarding playbook
  Who runs the calls? Proposal: Param for the five, with Sai or Arvind shadowing one each.
• Onboarding playbook
  When does the new Set-up tab reach production? Proposal: after the five partners' first calls, so they learn one screen at a time.
• Demo script
  Which stores should the demo show? Proposal: two or three public brands prepared by our team, never a design partner unless it says yes in writing.
• Demo script
  Does Arata stay on the demo? It agreed on 15 Oct, before it connected real accounts. Proposal: take it off the picker and use a public brand instead.
• Demo script
  Where does a live read of a prospect run? The guide says staging, which is now for testing only. Proposal: production, started by staff.
• Demo script
  Five minutes or ten? The guide is ten. Proposal: a five minute run, then five minutes of the founder's questions.
• Pricing and packaging
  v1 at $79 or $119? Proposal: $79, since v1 only drafts and hands off.
• Pricing and packaging
  Autopilot as a percent of spend, or flat fees by spend band? Proposal: 3% with the $149 minimum, as written.
• Pricing and packaging
  Who prices briefs to our team? Proposal: our team sets its own per brief price; Brew shows it, never bundles it.
• Pricing and packaging
  Does wallet money expire? Proposal: never, as written.
• Pricing and packaging
  Partner price for life? Proposal: yes, $79 or ₹4,999 for the five who sign the partner agreement.
• Pricing and packaging
  Is the spec committed? It sits untracked in the working tree. Proposal: commit it to brew main as a proposal so it cannot be lost.
• Contracts
  The whole document waits on you: Design partner agreement, order form
• Contracts
  Do we need a written agreement before the partnership goes further? Proposal: yes, the one page above, sent to all five this week.
• Contracts
  Which entity signs? Proposal: Yellow Monk Technologies LLP, the entity already on the Meta business and the site footer.
• Contracts
  What happens after the partnership? Proposal: accept the proposed partner price from pricing, or leave with nothing owed.
• Contracts
  Who reviews the terms? Proposal: a lawyer reads the agreement and the full terms once, before they replace the preview pages.
• Contracts
  Does the agreement cover paid makes from the wallet? Proposal: the wallet is a gift, unused money is not refunded, and it never expires during the partnership.

Customer success and support 40

• Partner onboarding
  Who on our team owns each partner's first week? Proposal: Param owns Arata and Nutrova, Sai and Arvind split the other three, one name per brand written here.
• Partner onboarding
  Is the first brief shared within 7 days the activation event? Proposal: yes for v1, and revisit once the five have each shared one.
• Partner onboarding
  Do invited teammates see the set up screens on first open? The code does not say clearly. Proposal: test it on Zenant with a fresh invite and record the answer here.
• Partner onboarding
  Should every partner get a welcome call in week one? Proposal: one 30 minute call on day 1 or 2, run from the checklist above.
• Help centre
  Where does the help centre live? Proposal: inside the product, from Help in the account menu, as one page with thirteen sections.
• Help centre
  Is the 7 day deletion promise ours to keep on production? Proposal: yes. The deletion tool, db/delete-brand.mjs, is on origin/main and reaches production with the next promote; until then an engineer does it by hand within 7 days.
• Help centre
  Do we publish the read limits by number? Proposal: say they exist and that a person can lift them, without numbers that can change.
• Help centre
  The read line promises 3 to 4 minutes. Proposal: measure ten first reads on production before it ships.
• Support process
  The whole document waits on you: Channels, response times, who answers
• Support process
  Who reads hello@joinsquareup.com each day? Proposal: Param on weekdays until the five partners are onboarded, then a rota of Param, Sai and Arvind.
• Support process
  What response times do we promise partners? Proposal: one working day for any question, same day for blockers and our mistakes, written into the partner welcome.
• Support process
  Do partners get a WhatsApp line? Proposal: not in v1; offer it in v2 once the inbox times hold for four weeks.
• Support process
  Do we need a ticket tool? Proposal: no; inbox labels plus Next are enough for five partners.
• Support process
  Who answers alerts at night? Proposal: nobody at night during the preview; the reader of the day checks alerts first thing.
• Known issues
  Should partners see a known issues page? Proposal: yes, a short one in the help centre with only what they can see, updated weekly.
• Known issues
  The interface still says Brew in a few places, now from one constant. Do we set that constant to the picked name, or to none, before the five partners use it daily? Proposal: keep Brew until the name is picked, then change the one line.
• Known issues
  Who decides the compute service account rights? Proposal: Param removes the project wide rights this week and gives each service its own account.
• Partner health
  The whole document waits on you: Signals that a partner is thriving or slipping
• Partner health
  Do we score health at all during the design partnership? Proposal: yes, a weekly green, amber or red line per partner, by hand, for the first four weeks.
• Partner health
  What counts as thriving? Proposal: an open in 7 days, one jam moved and every connection fresh, as written in Tomorrow.
• Partner health
  Who reviews it? Proposal: Param, every Monday, with each partner's owner.
• Partner health
  Does a partner ever see its own health? Proposal: no; it is our working view only.
• Feedback loop
  Where does all feedback live? Proposal: Next, with one row per report, linked from the QA sheet and reports.
• Feedback loop
  Who closes back to partners? Proposal: the partner's owner from partner onboarding.
• Feedback loop
  Do we send a weekly founder note? Proposal: yes during the design partnership, signed by Param, naming one thing that changed because of a partner.
• Notices to customers
  What counts as an incident worth a notice? Proposal: more than 30 minutes down, or anything that touches a brand's data or connections.
• Notices to customers
  Who sends notices? Proposal: Param during the design partnership, from hello@joinsquareup.com, signed with his first name.
• Notices to customers
  Do we pick a fixed maintenance window? Proposal: no fixed window; releases already roll back on their own, and planned downtime gets 2 working days notice.
• Notices to customers
  May Fetch appear in change notices? Proposal: only as the sign off line in the table, never in incidents, maintenance, connections or policy.
• Briefs sent to our team
  Who on our team picks up briefs? Proposal: Param until the first ten, then a rota.
• Briefs sent to our team
  What do we promise on time? Proposal: received within one working day, a first version or a question within three working days.
• Briefs sent to our team
  Do briefs to our team cost money during the design partnership? Proposal: free until pricing is unparked, then paid per brief as the 26 Oct pricing note says.
• Data requests
  When do we make the brew-prod-ops job? Proposal: right after the next promote, with one dry run on a test brand on staging first.
• Data requests
  The tool keeps no archive, only the audit row and the backups for up to 14 days. Do we say that on the privacy page? Proposal: yes, one line with the 14 days.
• Data requests
  What does a shared brief link show after its brand is deleted? Proposal: a plain page that says the brief was removed.
• Data requests
  Do we offer export in v1? Proposal: on request only, as a zip within 7 days.
• Every message
  Our company name is gone and the product name sits in one constant, still Brew. Do we keep Brew in the sender, header and Slack post until the name is picked? Proposal: yes, then change the one line.
• Every message
  May Fetch speak in emails at all, or only in the product? Proposal: only in the sign off of the Plays email and the first line of the digest, as in the table.
• Every message
  The read lines promise 3 to 4 minutes. Is that true? Proposal: measure ten first reads on production and use the measured time, or drop the number.
• Every message
  Should the sender be a person, such as Param at our team, rather than the product? Proposal: the product sends routine emails; anything personal comes from a named person.

Finance 49

• Pricing
  The whole document waits on you: The parked plan: free read, 14-day trial, plans and credits
• Pricing
  Accept Plan plus Wallet plus Autopilot as the model? Proposal: yes, as written on 26 Oct, recorded in the decisions log once Sai agrees.
• Pricing
  v1 at $79 or $119? Proposal: $79 or ₹4,999, since v1 only drafts and hands off.
• Pricing
  Autopilot as a percent or flat bands? Proposal: 3% with the $149 minimum.
• Pricing
  Wallet expiry. Proposal: never expires, as written.
• Pricing
  Partner price for life. Proposal: yes, $79 or ₹4,999 for partners who sign the partner agreement.
• Pricing
  Commit the spec. It sits untracked in the brew working tree. Proposal: commit it to brew main marked as a proposal.
• Unit economics
  What is a Sloosh credit in dollars? Proposal: Param adds the Pro plan's price and credit count to vendors and bills, and finals are reported in dollars from then on.
• Unit economics
  Which number do we plan on until a month is measured? Proposal: plan on $45 a brand a month, the v4 brain estimate, as a cautious middle; replace it with the measured figure in December.
• Spend caps
  Is $250 a day of first builds right for production? Proposal: keep it; five partners at $45 a build fit, and a new brand waits a day past that.
• Spend caps
  The build cap default in code is $60 but production runs $45. Proposal: set the code default to $45 so a missing env var never spends more.
• Spend caps
  A monthly ceiling per brand. Proposal: $50 a brand a month across all sources during the partnership, raised only by Param.
• Budget
  The whole document waits on you: Monthly spend by category
• Budget
  Accept the November ceilings? Proposal: yes as written; treg at $150 a month is well under its $40 a day cap and about fifteen times the $10 a brand estimate for seven brands.
• Budget
  Lower the treg day cap to match? Proposal: set TREG_DAY_CAP_USD to $15 on production, so the code cap and the budget agree.
• Budget
  Who approves spend over a ceiling? Proposal: Param only, by a yes in chat recorded in the spend ledger.
• Credits
  Top up Anthropic for production? Proposal: not during the partnership; Plays stay on Gemini Pro on credits until a fair quality rerun shows Opus is worth the cost.
• Credits
  Does the GFS credit cover Claude on Vertex? Proposal: ask Google for Startups in writing before any Claude on Vertex call.
• Credits
  Who reads the balances each month? Proposal: Sai, on the first Monday, with the spend ledger.
• Spend ledger
  The whole document waits on you: What was spent, on what, by month
• Spend ledger
  Where does the ledger live? Proposal: a sheet in the team's shared drive, linked here, with this page holding the monthly totals.
• Spend ledger
  Back fill October? Proposal: yes; Param adds the Sloosh, Apify, treg and Anthropic amounts from his card statements, and the rows above get their amounts.
• Spend ledger
  Who keeps it? Proposal: Sai, with Param approving each month.
• Revenue plan
  The whole document waits on you: Brands, plans, months to targets
• Revenue plan
  When does the first brand pay? Proposal: the month after the design partnership ends; target all five partners on the partner price.
• Revenue plan
  The first revenue target. No source sets one. Proposal: 25 paying brands within six months of the first charge, then reset from the real rate of new brands.
• Revenue plan
  Count wallet money as revenue when loaded or when spent? Proposal: when spent on a kept piece; the CA confirms the treatment.
• No spend while building
  Does the rule cover real reads for new prospects? Proposal: no; a real read for a prospect Param or Sai invited is product use, not building, and runs inside the caps.
• Vendors and bills
  Plan prices and renewal dates. Proposal: Param fills price and renewal for Sloosh, Apify, Resend, Cloudflare, ElevenLabs and the Anthropic top ups from his statements this week.
• Vendors and bills
  Which card or account pays each? Proposal: one company card for every vendor, in the LLP's name, so invoices match the entity.
• Vendors and bills
  Keep OpenAI? Proposal: keep the key as the last fallback in the media chain, with no top ups beyond what it holds.
• Financial model
  The whole document waits on you: Pricing, costs and growth tied together
• Financial model
  Growth inputs. Proposal: model three cases of 5, 15 and 30 new trials a month with a 30% trial to paid rate and 5% monthly churn, labelled projections, until the first real month replaces them.
• Financial model
  Cost per brand input. Proposal: use $45 until three partners have one measured month.
• Financial model
  Fixed costs and cash. Proposal: Param enters salaries, tools and cash privately in the sheet; this page shows only the runway.
• Financial model
  Who owns the sheet? Proposal: Sai builds it, Param reviews it monthly with the investor update.
• Billing and invoicing
  The whole document waits on you: How a brand pays: provider, invoices, plans in the product, failed payments
• Billing and invoicing
  Which provider? Proposal: Razorpay first, Stripe later, as above; the CA checks both for invoices that comply.
• Billing and invoicing
  Is the LLP registered for GST, and what is its GSTIN? No source records it. Proposal: Param confirms with the CA before the first invoice.
• Billing and invoicing
  Is wallet money billed with GST at top up or at use? Proposal: ask the CA; build whichever the CA confirms.
• Billing and invoicing
  Who can change a brand's plan? Proposal: the brand's owners and admins, and staff for partners.
• Taxes
  The whole document waits on you: GST and taxes in each market, invoices that comply
• Taxes
  Book the CA conversation. Proposal: Param books it this month with the six questions above and the pricing proposal attached.
• Taxes
  Show ₹ prices with or without GST? Proposal: show ₹4,999 plus GST to businesses, and say so on the plan sheet, once the CA confirms the rate.
• Taxes
  Dollar sales from India. Proposal: invoice in USD from the LLP under a letter of undertaking, if the CA confirms it applies.
• Refunds and collections
  The whole document waits on you: When we refund, how, dunning for failed payments
• Refunds and collections
  Accept the first month refund within 7 days? Proposal: yes, once per brand, so a new brand risks nothing.
• Refunds and collections
  Accept read only after 7 days of failed charges? Proposal: yes, never deleting, matching the trial's end.
• Refunds and collections
  Who sends the dunning emails? Proposal: the provider sends receipts; Brew sends the plain words emails from its own layout.

People and ways of working 30

• Team and roles
  The whole document waits on you: Param, Sai, Arvind: who owns what
• Team and roles
  Who owns what? Proposal, for Param to edit:
• Team and roles
  Param: company and strategy, product, brand, finance decisions, AI and models, ops and infrastructure, security and legal, this handbook.
• Team and roles
  Sai: sales and partnerships, customer success and support, design partner relationships, QA rounds through the QA sheet, and the money routines the finance pages propose: the spend ledger, the monthly credit read and the financial model; pricing is agreed by Param and Sai together.
• Team and roles
  Arvind: integrations and ecosystem, including the Shopify app and Built for Shopify, and marketing and growth.
• Team and roles
  Everyone: research and insights.
• Team and roles
  What are the titles? Proposal: none until a fourth person joins; describe people by what they own.
• Team and roles
  Who decides when Param is away? Proposal: Sai for partners and Arvind for integrations, with no production promotes until Param is back.
• How we work with Claude
  Do Sai and Arvind run Claude sessions on brew? No source says. Proposal: yes, each in their own worktree on a branch, merging through Param until they have done one QV round with him.
• How we work with Claude
  Which memory notes move into the handbook? Proposal: rules and runbooks move into handbook pages over the next two weeks; memory notes then keep a one-line pointer.
• Rituals
  Hold a weekly review? Proposal: yes, Mondays at 11 am IST, 30 minutes, agenda as above, starting 2 Nov 2026.
• Rituals
  How often do design partner audits run? Proposal: one full audit across the five partners every two weeks, run by Claude, triaged by Param and Sai.
• Hiring plan
  The whole document waits on you: Roles, when, why
• Hiring plan
  What triggers the first hire? Proposal: when two design partners agree to pay, or when Param owns more than 30 open items in Next for two weeks running.
• Hiring plan
  What is the first role? Proposal: a senior full-stack engineer who can run production: promote, restore, on call, and review Claude sessions' work. It removes the single point of failure.
• Hiring plan
  What comes second? Proposal: a design partner success person, once there are more than eight brands.
• Hiring plan
  Contractors before hires? Proposal: yes for illustration of the picked brand direction; see Contractors and agencies.
• Onboarding a teammate
  The whole document waits on you: Day one to week two
• Onboarding a teammate
  Accept the proposed path? Proposal: accept as written, and try it on Sai or Arvind if either has not yet run brew locally.
• Onboarding a teammate
  Who is the buddy for a new person? Proposal: Param for engineers, Sai for anyone working with partners.
• Onboarding a teammate
  When does a new person get production access? Proposal: after 30 days and one restore test with Param.
• Tools we use
  Who has access to what? No source records it. Proposal: Param fills one access column per person this week; until then, assume only Param has admin on Google Cloud, Cloudflare, GitHub, Meta, Shopify Partners and Slack.
• Tools we use
  Move personal logins to team accounts? treg and GitHub sit on personal accounts in our notes. Proposal: move treg, Cloudflare and GitHub to accounts on joinsquareup.com, with Param and one other admin each.
• Tools we use
  Use a password manager? Proposal: yes, a team vault holding every login above, with Sai and Arvind invited.
• Changing this handbook
  Who besides Param may set a document to verified? Proposal: only Param, until each department has a named owner; then the owner may verify their own department.
• Changing this handbook
  When is Handbook 1.0 taken? No snapshot exists yet. Proposal: the day every department is written and check.mjs passes, before the first upload.
• Contractors and agencies
  The whole document waits on you: Who we work with outside the team, for what, on what terms
• Contractors and agencies
  Hire an illustrator once a direction is picked? Proposal: yes, a fixed-price brief for the character sheet and 12 key poses, reviewed by Param and Sai.
• Contractors and agencies
  Who signs contracts? Proposal: Param, for the company, on a one-page template kept in Security, privacy and legal.
• Contractors and agencies
  May a contractor run Claude sessions on brew? Proposal: no, until they are onboarded like a teammate.

Product marketing 55

• Positioning
  The whole document waits on you: What we are, for whom, against what, the one line
• Positioning
  The final positioning statement. Proposal, to accept as is: "For founders of online stores who are their own marketing team, the product is your store's marketing team, in one feed. It reads your store and everything said around it, and drafts this week's marketing with the proof beside every number, for you to pick and hand to whoever ships. Unlike a chat assistant or a design app, it starts from your own store, not a blank box." Fetch closes it with: This week's ideas, fetched.
• Positioning
  Tool, team or service? Proposal: team in the line, tool in search terms; test the say back in five partner calls before the site goes up.
• Positioning
  Do we say AI in the line? Proposal: no; open with the store, and say AI in the body copy.
• Positioning
  Does the line name a country? Proposal: no; the product is global, and India first is a go to market choice, not a position.
• Positioning
  Does team mislead, since our team also ships briefs? Proposal: keep it; our team appears only as one of the people you can hand off to, priced per brief.
• Messaging framework
  Is the roof line final? Proposal: yes, Your store's marketing team, in one feed, the line already live.
• Messaging framework
  Do the three jobs become public words? Proposal: yes in the product and the docs as how picks are grouped; keep the four pillars as the message.
• Messaging framework
  Which of Fetch's hooks leads? Proposal: the founder hook on Instagram first: You throw the link. It brings back the week.
• Messaging framework
  Does Fetch's own pillar wording replace the shared one? Proposal: no; the shared pillars stay, and Fetch's wording lives only in its posts and reels.
• Value proposition by persona
  Which persona does every public page lead with? Proposal: the founder who is the marketing team; the others get their own page or slide.
• Value proposition by persona
  Do we promise anything to agencies before the brand switcher? Proposal: only the brief written for them, and say plainly that each client is its own invite.
• Value proposition by persona
  Do we keep the ten minutes as a public line? Proposal: no, until it is measured; say the week, already drafted, instead.
• Competitive intelligence
  Which named tools do we research first? No source names any. Proposal: the first three that design partners name in calls, nothing before.
• Competitive intelligence
  May we spend on competitor research? Proposal: free pages and the Ad Library only until the partnership ends.
• Competitive intelligence
  Do we ever publish a comparison page? Proposal: no; compare jobs in public and offer a read instead.
• Competitive intelligence
  Is our team a competitor to agencies, since it ships briefs? Proposal: no; say our team is one of the people a brand can hand off to, priced per brief, never bundled.
• Launch tiers and playbooks
  Is there a public launch day at all? Proposal: no single day; open reads in small batches and treat v2 as the first tier one launch.
• Launch tiers and playbooks
  Who calls the tier? Proposal: Param, at planning, in the release's first line.
• Launch tiers and playbooks
  Does the name come before any launch? Proposal: yes; no tier one or two launch goes public before the name and the direction are picked.
• Launch tiers and playbooks
  When do reads open past the five partners? Proposal: when two partners hand off weekly for three weeks running, the same bar as v2 planning.
• Sales enablement kit
  Where do the one pager and slides live? Proposal: one published deck and one page in the chosen direction, made once the direction is picked; until then, the spoken version and the demo only.
• Sales enablement kit
  Does Fetch appear in an investor pitch? Proposal: on the cover and end card only; the story is the product.
• Sales enablement kit
  Do we show a partner's read on slide 3? Proposal: only with that partner's written yes; otherwise the sample candle store, labelled Sample.
• Sales enablement kit
  Who may send the leave behind? Proposal: Param, Sai or Arvind, from their own email.
• Objections and answers
  Do drafts or reads train anyone's AI model? No source answers it. Proposal: check the Vertex terms for our project, then add one line to the privacy page and a row here.
• Objections and answers
  How long do we keep data after a brand leaves? No source says. Proposal: 90 days, then delete, written into the privacy page.
• Objections and answers
  What do we say to a store that is not a fit? Proposal: say so plainly and offer the demo stores instead of a read.
• Objections and answers
  Do we name Gemini in public? Proposal: yes, as Google Cloud's AI, since the privacy page already says AI models on Google Cloud.
• Claims we can make
  Do we retire about 160 cards for good? Proposal: yes; quote no card count until a partner's feed is counted on the day.
• Claims we can make
  Do we say 60 sources in public? Proposal: yes, as "reads 60 sources today", since the inventory backs it and it is checked by the docs truth check.
• Claims we can make
  Do we name the AI tools we check? Proposal: yes, ChatGPT, Gemini, Perplexity and Claude, since the weekly check asks them; under Gemini the Claude ask is skipped, so check the setting before a campaign.
• Claims we can make
  May we name the partner behind 12 of 32? Proposal: no; keep it as an example finding from one brand's read until that partner says yes in writing.
• Claims we can make
  Who fixes the three places that break the rules? Proposal: one product change in the next release, checked on staging.
• Customer stories and proof
  The whole document waits on you: Partner results we can share, with permission
• Customer stories and proof
  Which partner do we ask first? Proposal: Arata, since it is furthest along and has real accounts connected, after its next weekly hand off.
• Customer stories and proof
  What is the first story, before results exist? Proposal: a finding from their read, with their yes, told as what we found, never as a result.
• Customer stories and proof
  Who asks? Proposal: Param, in a personal email, never a form.
• Customer stories and proof
  Is Gully Labs still a partner we may ask? Proposal: treat it as a past prospect and do not ask until Param confirms.
• Pricing and packaging story
  Is "wallet" the word brands see? Proposal: yes; it says real money and avoids credits.
• Pricing and packaging story
  Do we show $ and ₹ side by side? Proposal: no; show the visitor's market only, ₹ in India and $ elsewhere.
• Pricing and packaging story
  Do we tell partners their price for life now? Proposal: not until Param and Sai agree the proposal; then in a personal note, before any public page.
• Pricing and packaging story
  Do the plan's internal words, the brain and the hands, appear in public? Proposal: no; say what it thinks and what it makes.
• Win and loss
  The whole document waits on you: Why brands say yes or no, from every conversation
• Win and loss
  What did each of the five partners say when they agreed? No source records it. Proposal: Param writes one record for each from memory, marked as recalled, this week.
• Win and loss
  Why did Gully Labs drop off the list? Proposal: Param writes its record as paused or lost with the reason it gave.
• Win and loss
  Where do records live? Proposal: on this page while there are fewer than twenty; move to the pipeline sheet after.
• Win and loss
  Who asks the three questions? Proposal: whoever had the last call, within a week.
• Product pages and explainers
  Which domain do the pages live on? Proposal: decide it with the name; keep feed.joinsquareup.com as the app.
• Product pages and explainers
  One page or two, home and ask? Proposal: one page, with the ask as its last block, and the Full state on the same page.
• Product pages and explainers
  Where do explainers live? Proposal: on the home page under each block, and the same cuts on Instagram.
• Product pages and explainers
  What replaces the landing's unlabelled sample cards? Proposal: label them Sample now, and swap in a partner's finding only with written yes.
• How AI answers describe us
  Do we run a check before the name is picked? Proposal: no; run only the unbranded questions once, free in a chat by hand, to learn which tools get named.
• How AI answers describe us
  Which market's answers do we check? The brand check asks from India. Proposal: India first, since every partner is Indian, and add the US when we open past India.
• How AI answers describe us
  Who reads the answers each week? Proposal: Sai, ten minutes on Monday, and Param approves any page fix.

Research and insights 24

• Research plan
  Which three questions go first? Proposal: 1, 2 and 3, all on the weekly call slot.
• Research plan
  May design partners be asked to test, and how often? Proposal: one 30 minute session each in November, inside the weekly call.
• Research plan
  Who runs the calls? Proposal: Param runs them, Sai or Arvind takes notes into the library the same day.
• Research plan
  Run the model bake off? Proposal: first the fair rerun of five Plays on Pro, about $1, then the bake off, about $5 to $10.
• Research plan
  Is India first? Proposal: run the India counts first and keep global open until they are in.
• Interview and call library
  Have there been partner calls that were never written down? Proposal: Param lists each one he remembers with the five and Gully Labs, with a date, and we add them marked "logged from memory".
• Interview and call library
  Do we record calls? Proposal: yes, screen and voice with consent, deleted once the notes are in this library.
• Interview and call library
  Is APE's voice interviewer used for Brew research? Proposal: no, not in v1; talk to partners in person on the weekly call.
• Voice of the customer
  May we quote partners by name inside the handbook? Proposal: yes, inside only; never outside without written permission.
• Voice of the customer
  Who reads card notes each week? Proposal: Sai, every Monday, for the five partners.
• Voice of the customer
  Should Param's testing keep its own section? Proposal: yes, labelled our team, until partner words outnumber it.
• Usability tests
  The whole document waits on you: Tests run, tasks, findings, fixes
• Usability tests
  Do we run the first round? Proposal: yes, one person from each of Arata, Nutrova, Mana, Puma India and Scrubroom, in the two weeks from 2 Nov, inside the weekly call slot.
• Usability tests
  Live build or the redesign screens? Proposal: the live build first, so the redesign has a baseline to beat.
• Usability tests
  Which scripts first? Proposal: the instant test plus script 1 for every person, script 2 a week later.
• Usability tests
  What is the pass bar? Proposal: four of five people pass all four instant signals.
• Usability tests
  Who moderates? Proposal: Sai moderates, Param watches with the camera off and says nothing.
• Market and category research
  Is India first? Proposal: yes for counting; size India first and keep global open.
• Market and category research
  Do we pay for a store technology lookup? Proposal: not yet; try free platform figures and a hand sample of 100 stores first.
• Market and category research
  Who checks each number? Proposal: whoever did not gather it, Param, Sai or Arvind.
• Market and category research
  Do large brands like Puma India count in our ring? Proposal: keep them as partners to learn from, size only founder led stores.
• Insight repository
  Should insights from QA count here, or only from people? Proposal: both, kept in separate sections, so the people section shows how empty it is.
• Insight repository
  Run the fair model rerun behind insight 10? Proposal: yes, five Plays on Pro with today's inputs, about $1.
• Insight repository
  Who owns this page? Proposal: Param, with Sai adding rows after each QA round.

PR and communications 30

• Press kit
  Who are the founders on the kit, and what are their roles? No source gives roles for Param, Sai and Arvind. Proposal: Param writes three bios of 40 words each, with one photo each, before any kit goes out.
• Press kit
  Do we name Yellow Monk Technologies LLP in the kit? Proposal: yes, in the footer line only, as the joinsquareup.com footer already does.
• Press kit
  Can any design partner be named? Proposal: no until one gives a written yes; say "a small group of real brands".
• Press kit
  Which screenshots? Proposal: new ones on the sample candle store once the name is picked, never a design partner's feed.
• Press kit
  Which tagline goes on the kit? The direction data says "This week's ideas, fetched." and the old book said "This week's ideas, dropped at your feet." Proposal: the direction data line. [src: brand-directions/docs2/data/dirs.json] [src: brand-directions/book/content/h.js:19]
• Announcements calendar
  Is there a public launch day at all? Proposal: no single day; open reads in small batches, as the roadmap proposes.
• Announcements calendar
  Which channel opens first? Proposal: Param's LinkedIn for the founder post and Instagram for the character, as the old six week plan assumed.
• Announcements calendar
  Do we build a waitlist before announcing anything? Proposal: yes, a one field form on the landing, before the name announcement.
• Announcements calendar
  Is the Shopify listing worth a public announcement? Proposal: yes, because it is the one way in that needs no setup.
• Founder voice
  The whole document waits on you: What Param writes and posts, the topics, the tone
• Founder voice
  Which channel? Proposal: LinkedIn first, as the old plan assumed.
• Founder voice
  Who writes, Param alone or all three founders? Proposal: Param posts weekly; Sai and Arvind once a month each, on their own work.
• Founder voice
  Which topics first? Proposal: the six above, in that order, starting the week the name is announced.
• Founder voice
  Does the weekly note to partners start now? Proposal: yes, every Friday from 30 Oct, from Param, with no character.
• Founder voice
  What tone? Proposal: plain, specific and honest about what the product does not do yet, the way the claims page reads.
• Media list
  The whole document waits on you: Journalists, newsletters and podcasts worth reaching
• Media list
  Who are the first five names? Proposal: Param lists five people the team already reads or knows, India first, and this page records them with why.
• Media list
  India first, or global? Proposal: India first, since every partner is Indian; copy stays plain global English.
• Media list
  Who owns outreach? Proposal: Param, with one named backup from Sai or Arvind.
• Media list
  When do we start? Proposal: the week after the name is announced, with the press kit attached.
• Crisis communications
  The whole document waits on you: What we say when something goes wrong in public
• Crisis communications
  Who speaks in a level 3 crisis? Proposal: Param only; Sai and Arvind forward every question to him.
• Crisis communications
  How fast is the first word? Proposal: within one hour for level 3, the same day for level 2, as the comms books say.
• Crisis communications
  Do we keep a public status page? Proposal: no in the preview; email each affected brand instead.
• Crisis communications
  What do we say if Meta or Shopify suspends the app? Proposal: tell partners the same day which connector stops, that their data is safe, and that the other connectors keep reading.
• Crisis communications
  Who is the legal contact for a data exposure? No source names one. Proposal: Param names one this week.
• Internal updates
  Do we want a weekly note at all, or is the morning digest enough? Proposal: a weekly note on Mondays, because Sai and Arvind do not see the digest.
• Internal updates
  Where does it go? Proposal: email from Param to Sai and Arvind, with a link to the handbook snapshot.
• Internal updates
  A weekly meeting? Proposal: 30 minutes on Mondays after the note, to answer the three questions.
• Internal updates
  When is Handbook 1.0? Proposal: when every document in every department is written.

Integrations and ecosystem 29

• Shopify app
  What did Shopify answer? No outcome is recorded after 21 Oct. Proposal: Param checks hello@ and the Partner Dashboard today and notes the state here.
• Shopify app
  The repo draft now says Brew: Daily Marketing Plays; the listing was submitted as Brew: Daily Marketing Moves. Proposal: change the live listing to the Plays wording only after approval, so the review is not reset, then rename to the picked name.
• Shopify app
  Fully visible at once, or limited to the link? Proposal: stay limited to the direct link until the five design partners have installed it and production carries the app.
• Shopify app
  Webhook version 2026-10 against Admin API 2026-04. Proposal: keep both, and bump the Admin API to the newer version in one tested change after approval.
• Shopify app
  Ask for read_all_orders? Proposal: no; 60 days of orders is enough for every Play today.
• Meta
  Has Business Verification passed since 24 Oct? No later source. Proposal: Param checks today and notes the date here.
• Meta
  pages_show_list is granted at sign in but not in the review list. Proposal: drop it from the login configuration before Submit, since the Page is now found from the ads.
• Meta
  Who records the screencasts? Proposal: Param, on Cozy, the same day verification passes.
• Meta
  The old Harry and Donna datasets still take events. Proposal: switch off their pixel and CAPI senders at the source so the portfolio carries only live products.
• Google
  Did Ads Basic access come through? No source after 16 Oct. Proposal: Param checks the API Center today; until then the developer token keeps reading as it does.
• Google
  Search Console as a connector. Treg's catalog lists it as connected and free, but no Brew connector reads it [src: memory treg-catalog-for-brew]. Proposal: add it in v2 together with any other new scope, in one review.
• Google
  One project for Brew and the old cockpit. Proposal: keep sharing it, because the developer token is tied to this project, and retire the cockpit's Google connect instead.
• Slack
  List the app in the Slack Marketplace? Proposal: no until the name is picked; the listing would carry the old name.
• Slack
  Who may press Approve? Param chose anyone in the channel on 17 Oct. Proposal: keep it, and record who pressed it in the jam as it does now.
• Slack
  The manifest now says plays, and still names the app Brew. Proposal: push the new description with tools/slack-app.mjs update now, and rename the app in one more update once the name is picked.
• Data and making partners
  Who is our contact at each partner? Proposal: Param names one person per partner and the email they answer, and this page lists them.
• Data and making partners
  Put the Higgsfield keys in Secret Manager now? Proposal: yes, on staging only, with the AI_MOCK tests passing first; production after one paid Studio make on a test brand with Param's yes.
• Data and making partners
  Sloosh on production. Proposal: keep it off until a design partner asks for a Sloosh final; then sign in on production and turn SLOOSH_CHECK on.
• Data and making partners
  Talk to Higgsfield sales about Enterprise? Proposal: yes, one call, to ask whether many brands on one account is allowed for the CLI only features.
• Data and making partners
  Which treg account owns the bill? Treg runs on Param's own sign in for team squareup. Proposal: move the team's billing to a company email.
• API and MCP for agencies
  The whole document waits on you: How an agency or AI tool could plug in
• API and MCP for agencies
  Do we build an agency login before any API? Proposal: yes; agencies are the people briefs go to, and the agency route is already one of the five open calls.
• API and MCP for agencies
  Should Brew offer its own MCP server? Proposal: yes in v2, read only, one brand per token, starting with Plays, briefs and the Brand Book.
• API and MCP for agencies
  Who is the first user of it? Proposal: one design partner's agency, named by Param, before any public docs.
• API and MCP for agencies
  Is it paid? Proposal: included in the plan while pricing is parked; decide with pricing.
• API and MCP for agencies
  Does an agency see connected numbers? Proposal: no, unless the brand's owner says yes per agency.
• Integration roadmap
  Which comes first after the approvals: marketplaces or Instagram reads? Proposal: own Instagram reads, because they are free and every brand has an account.
• Integration roadmap
  Do write scopes start at v2 or v3? Proposal: v2 for posting only, as the version plan says, and fix the engineering table to match.
• Integration roadmap
  Do we build email or WhatsApp sending at all? Proposal: no in v1 and v2; send through the brand's own Klaviyo or WhatsApp tool.