One agent
$79₹4,999per month
Pick any one of the sixteen.
- Runs on its schedule, hosted by us
- Approvals in Telegram
- The brain, built and maintained
- Monday brief included
For Shopify owners and the performance marketers they hire
A marketing brain that acts.
Sixteen agents read one versioned brain about your store, run on a schedule, and stop for your tap before anything spends, sends or changes the store. Then they write what they learned back.
It runs. It sends. You read.
- What it is: a brain document plus sixteen agents with written contracts.
- What it never does: spend, send or write to your store without an approval you tapped.
- What it costs: $79 a month for one agent, $249 for four, $549 for all sixteen.
- How to try it: Competitor Watch and the Weekly Brain Update free for 14 days, no card.
- Every campaign created paused, verified by status read back
- One use approval tokens that expire
- Every store change ships with its previous value
- Learnings append only, dated, sourced
StoreCrewbot
Store Fix Approval
Apply 7 content changes to 4 live products? Here are three of them.
- Product description, hero product 1Premium quality. Order now.Uses the phrase your reviews repeat most. Names the objection your FAQ never answered.
- SEO title, hero product 2Product 2Product 2 in the words customers search for, under 60 characters.
- FAQ metafield, hero product 3emptyQuestion taken verbatim from customer.objections, answer inside claims_allowed.
Previous values recorded. One live batch per day. Token expires in 24 hours.
ApproveHold
The closed loop, in five beats.
This is the mechanism behind the claim. Other marketing brains stop at beat one.
- The brain is read
An agent opens the brain before it works: identity, customers, products, competitors, performance, learnings. Store Fix reads your objections and allowed claims. Ad Creative reads what worked.
Brain document, six layers, version number visible.
- The crew does the job
Each agent follows a written workflow with quality gates. Every finding cites its evidence. A draft that would fit any store in your category is rejected and redrafted against your customers' vocabulary.
Agent run, steps ticking, evidence URLs attached.
- Your tap, before the risky part
Anything that spends, sends or writes to the store produces an approval card in Telegram: what, how many, three samples. Approve or hold. The token is single use and expires. No record means no action.
Telegram approval card with two buttons.
- The approved payload runs, nothing else
On approve, the agent executes exactly the payload you saw. Store content changes return their previous values. Ads stay paused until a separate approval activates them.
Changes file with before and after per field.
- What it learned goes back
Dated, sourced, append only: what worked, what did not, the decision you made and what happened after. Monday's brief is built from it. The brain gets smarter instead of staying a document.
Brain version increments, learnings line appended.
Brain, crew, approvals, results, back to the brain. One product, run from Telegram.
How a week runs.
The hosted crew keeps this rhythm on your local time. You read Telegram. Times are the agents' scheduled runs.
- Monday
- 06:00 Weekly Brain UpdateThree lines: the headline number and its delta, the most important thing that happened, the first thing to do. Under 60 words, one link to the full brief.
- 06:00 Competitor WatchDigest of what changed at your three to five competitors. Maximum seven items, each with the URL where it was observed. A quiet week is reported as a quiet week.
- 08:00 Content PublishingPosts for the coming days queued for your approval.
- Morning Store FixSunday night's re audit is waiting: fresh score, fresh queue, drafted fixes, one approval for the live batch.
- Tuesday
- 07:00 Creator OutreachTen new candidates qualified on real engagement. First emails drafted. Budget exposure shown as a number before the send batch asks for your tap.
- Wednesday
- 06:00 Retention and WinbackWho is about to lapse, which products are due for replenishment, campaign drafts held for approval.
- 07:00 UGC SeedingA new batch of micro creators who take product for content. Offers pause for approval.
- Thursday
- 08:00 AffiliateRecruitment batch for approval, codes and links for accepted affiliates, attributed orders monitored.
- 08:00 Content PublishingSecond queue of the week.
- Friday
- All day The daily agentsAd Growth at 07:00 and Reviews at 09:00 as every day. Nothing else scheduled. You read.
- Saturday
- All day The daily agentsSame. Approvals wait for you. Tokens expire rather than execute.
- Sunday
- 22:00 Store FixWeekly re audit of the storefront against the checkpoint library. The queue is waiting Monday.
Every day
- Daily 07:00 Ad GrowthReads the account, pauses rule fired losers (pausing costs nothing), asks before any budget goes up.
- Daily 09:00 ReviewsDrafts replies to 4 and 5 star reviews for the day's batch. Anything under 3 stars goes to you untouched.
- Every 30 minutes Post-PurchaseNew orders get the pre approved template through your connected channel. Risk flags reach you as a digest, never as a cancellation.
- Every hour Social EngagementComments and DMs drafted from the brain. A complaint reaches a human within the hour.
Monthly
- 1st and 15th, 05:00 Ad CreativeTen fresh creatives against what worked, uploaded paused. Activation waits for your tap.
- 1st of the month, 04:00 SEO and AI SearchTechnical audit, keyword targets, AI visibility sweep across ChatGPT, Gemini, Perplexity and AI Overviews. Fixes land as drafts.
Weekly Brain Update
Three lines: the headline number and its delta, the most important thing that happened, the first thing to do. Under 60 words, one link to the full brief.
06:00Competitor Watch
Digest of what changed at your three to five competitors. Maximum seven items, each with the URL where it was observed. A quiet week is reported as a quiet week.
06:00Content Publishing
Posts for the coming days queued for your approval.
08:00StoreCrewbot
Store Fix Approval
Apply 7 content changes to 4 live products? Here are three of them.
- Product description, hero product 1Premium quality. Order now.Uses the phrase your reviews repeat most. Names the objection your FAQ never answered.
- SEO title, hero product 2Product 2Product 2 in the words customers search for, under 60 characters.
- FAQ metafield, hero product 3emptyQuestion taken verbatim from customer.objections, answer inside claims_allowed.
Previous values recorded. One live batch per day. Token expires in 24 hours.
ApproveHold
Monday briefWeekly Brain Update
- HeadlineRevenue last week, against the week before.
- What happenedThe most important move and why, written as likely because with the evidence, or no clear cause.
- Do firstOne action, ranked by expected impact, naming the agent that does it.
Based on: the data you connected or pastedFull brief: one link
Every card in Telegram shows the agent, the count, three samples and two buttons. WhatsApp follows. A web dashboard holds the full outputs.
The architecture in plain words.
Six parts. You can point at each one on the diagram and ask what it does.
Operator in the middle. Brain on the left, read and written by the crew. Approvals gate the connectors on the right. Caps sit under everything.
- 1. The brain
One versioned document per store with six layers. Every agent reads it first. Only declared sections can be written by each agent, and only the Weekly Brain Update writes performance. Learnings are append only.
- 2. The Operator
The chat surface in Telegram. It collects the inputs an agent needs before a run is dispatched, delivers the outputs, and carries the approval cards. It never guesses a missing input; it asks.
- 3. The crew
Sixteen agents, each with a written contract: job, inputs, brain reads and writes, workflow, modes, output contract, quality gates, guardrails. Agents never call each other. The brain is the only shared state.
- 4. Approvals
A record with a single use, expiring, signed token. Tampered, wrong or unknown tokens are ignored. A replay is refused. On approve, the agent executes exactly the payload that was shown and nothing else.
- 5. Connectors
Shopify through GraphQL on a custom distribution app with named scopes. Meta Ads through the official API, paused by default. Klaviyo or Brevo, your mailbox, Judge.me, Zernio, WhatsApp Cloud API, Instagram Login API. Official APIs only for messaging. No unofficial libraries, ever.
- 6. Caps and the ledger
Every run checks your daily and monthly spend cap before it starts. Over cap means queued with a reason and an alert, never execute first and bill later. Every change is logged with the rule or approval that caused it, under your customer id only.
- 1. The brain
One versioned document per store with six layers. Every agent reads it first. Only declared sections can be written by each agent, and only the Weekly Brain Update writes performance. Learnings are append only.
- 2. The Operator
The chat surface in Telegram. It collects the inputs an agent needs before a run is dispatched, delivers the outputs, and carries the approval cards. It never guesses a missing input; it asks.
- 3. The crew
Sixteen agents, each with a written contract: job, inputs, brain reads and writes, workflow, modes, output contract, quality gates, guardrails. Agents never call each other. The brain is the only shared state.
- 4. Approvals
A record with a single use, expiring, signed token. Tampered, wrong or unknown tokens are ignored. A replay is refused. On approve, the agent executes exactly the payload that was shown and nothing else.
- 5. Connectors
Shopify through GraphQL on a custom distribution app with named scopes. Meta Ads through the official API, paused by default. Klaviyo or Brevo, your mailbox, Judge.me, Zernio, WhatsApp Cloud API, Instagram Login API. Official APIs only for messaging. No unofficial libraries, ever.
- 6. Caps and the ledger
Every run checks your daily and monthly spend cap before it starts. Over cap means queued with a reason and an alert, never execute first and bill later. Every change is logged with the rule or approval that caused it, under your customer id only.
Install link to first approval.
Your part takes under ten minutes. The crew works while you get on with the store.
- Click.
Open the install link
We email it. It installs a custom distribution Shopify app, so there is no App Store listing and no review queue. Not on Shopify? Send your store URL and start at level 1.
- Read the list. Approve.
Approve the scopes
Shopify shows you exactly what the app can read and write. Read products, write products (descriptions, SEO fields, FAQ metafields), write content (draft pages). Nothing on price, inventory, checkout or customers.
- Type six answers.
Answer six questions in Telegram
What you sell. Who buys. Your price band. Three competitors. What customers love. What they complain about. If you skip one, the Brain Builder fills it from storefront evidence and marks it inferred.
- Nothing. Wait a few minutes.
The brain is built
The Brain Builder crawls your storefront and up to 200 public reviews, lifts the ten most repeated customer phrases verbatim, classifies your category, and writes version 1. You get one screen: what I found, what I inferred, three things that look wrong.
- Reply once.
Correct it
Reply with corrections. The brain updates to version 2. Every claim in it maps to a source URL or an interview answer.
- Nothing.
Store Fix runs first
It fetches your home, top product pages, collection, cart and policy pages at desktop and 390px mobile, scores every observable checkpoint 0, 1 or 2 with a URL as evidence, and ranks by a formula you can check: (2 minus score) times impact weight times confidence, divided by effort.
- Tap approve or hold.
Your first approval card lands
Apply n content changes to k live products? Here are three of them, before and after. Previous values recorded. One live batch per day.
- Read.
Monday, the brief
Three lines under 60 words. Until you connect orders or paste last week's revenue and ad spend, it says so plainly: I do not have your numbers yet.
Left column: what you do at each step.
The sixteen, with their contracts open.
Every agent page shows what it reads from the brain, its workflow step by step, the tools it uses, what changes across Advise, Act and Autonomous, what waits for your tap, what it writes back, and an example of its output.
- Store Fix Finds what is losing sales on your store, ranks it by a formula you can check, writes the fixes in your voice, and applies the copy ones after one tap. Core Four default
- Ad Creative Turns one product page and the brain into a batch of finished static ads with a claims ledger and a QA sheet, then uploads them paused. It never unpauses and never sets a budget. Core Four default
- Ad Growth Reads your ad account every morning, pauses losers and scales winners on written, versioned rules, and asks before any budget goes up. It turns Ad Creative's batch into a managed account. Core Four defaultContract level
- Creator Outreach Finds creators in your niche, qualifies them on real engagement, writes a first email that cites one of their actual posts, negotiates inside your budget band, and chases the deliverable. Email first, always. Core Four default
- Competitor Watch Watches three to five named competitors every week and tells you only what changed and what you could do about it. Read only by design. One of the two free trial agents. Free 14 day trial
- Weekly Brain Update Every Monday it reads what happened last week across the store, the ads and the email, writes the learnings into the brain, and sends you three lines with the three things to do this week. It is the closed loop. One of the two free trial agents. Free 14 day trial
- Lifecycle Writes your seven always on flows in your voice and builds them in your email or WhatsApp tool, disabled, so you enable each one with a tap. Contract level
- Post-Purchase Watches new orders and sends delivery updates, review requests and replenishment nudges through your connected channel. Where COD exists it confirms the order and scores return risk, then flags risky orders to you. It never cancels anything. Contract level
- Retention and Winback Segments your customers by recency, frequency and value, finds who is about to lapse, predicts when each product runs out from real order intervals, and drafts the winback, loyalty and replenishment campaigns. This is where repeat revenue lives. Contract level
- Reviews Grows reviews by timing requests to after delivery and running disclosed incentives inside platform policy. Responds in your voice to 4 and 5 stars after you approve the day's batch, and routes everything under 3 stars to a human. Feeds recurring objections back into the brain. Contract level
- SEO and AI Search Seven jobs: keyword research, technical audit, product page SEO, programmatic page drafts, a monthly AI visibility sweep across ChatGPT, Gemini, Perplexity and AI Overviews, citation ready content, and the monthly organic calendar that Content Publishing executes. Contract level
- Content Publishing Researches what is working in your category, writes 30 days of reels scripts, carousels and captions in your voice, renders the carousels, and schedules them through Zernio after your approval. Contract level
- Social Engagement Drafts replies to comments and DMs in your voice, answers product questions from the brain, routes purchase intent, flags complaints to a human within the hour, and listens for mentions. Instagram Login API only. There is no compliant cold DM, so it does not do one. Contract level
- UGC Seeding Finds micro creators who take product for content, sends the offer from your mailbox, creates the Shopify draft order at 100 percent discount, tracks delivery, chases once, collects usage rights, and files the content where Ad Creative can use it. Contract level
- Affiliate Finds affiliate candidates, recruits them by email, creates unique discount codes and tracked links, monitors attributed orders, flags top performers, and prepares the monthly payout report. Payouts are a report you pay, never a transfer this agent makes. Contract level
- Offer and Landing Page Builds campaign landing pages as drafts and tests offers: bundle versus discount versus gift. Creates the codes, points Ad Creative's batch at the right page, and reads the result. No autonomous mode on purpose. Contract level
The Brain Builder is included with every plan. It writes version 1 of the brain and has one mode. It is part of the brain, not one of the sixteen.
Eleven of the sixteen are at contract level: their workflow is fully specified for Advise mode and the tool wiring, and the product owner expands Act mode before it ships. Their pages mark anything not yet in the contract as Planned.
Guardrails, and what the crew will never do.
Taken from the agents' own contracts. If a line here is not on an agent page, tell us and we will fix the page.
- Fail closed
Any action that spends, sends or writes to the store needs an approval record with a valid, unexpired, unused token. No record means no action. An expired approval means no action.
- Caps before dispatch
Daily and monthly spend caps are checked before a run starts. Over cap means queued with a reason and an alert.
- Paused by default
Every Meta campaign, ad set and ad is created paused. The agent reads back the status and fails the upload if it is not paused. Activation is a separate approved action.
- Claims from your storefront only
claims_allowed is a subset of text found on your own store. Anything outside it is rewritten or removed and listed as needs your confirmation.
- Evidence or it is cut
Every finding cites a URL or a run output. Uncited recommendations are removed. Causation is written as likely because with evidence, or no clear cause.
- The gate never relaxes
Autonomous mode changes the schedule, not the approvals. First emails, activations, budgets, live content and publishes wait for a tap in every mode.
The crew never
- spends money without an approval you tapped.
- unpauses an ad or sets a budget except through the approval gated tool.
- edits a campaign it did not create.
- touches price, inventory, checkout, shipping, tax or customer data when fixing your store.
- sends a cold Instagram DM. First touch to a creator is always email.
- sends outreach from our domain. Only from your connected mailbox.
- a fourth message to a creator who stayed silent.
- promises a creator a fee above your budget band. The tool refuses.
- gates negative reviews. Never replies under 3 stars without a human.
- offers a refund or discount in public.
- cancels an order. It flags and holds.
- makes a payout. Affiliate payouts are a report you pay.
- logs in to, submits a form on, or contacts a competitor.
- stores Google review text.
- says ROAS without blended, and never repeats platform reported ROAS.
- posts to X unless you switch it on. It costs $0.20 per post with a link.
How you undo
Store content changes ship with every previous value. Ads live in a campaign named StoreCrew plus the date, paused. Flows are created disabled. Pages are created as drafts. Any agent can be paused. Plans are monthly.
Three data levels. You choose how far.
Every agent states which level it ran at and, once, what a higher level would add. Then it stops selling.
-
Storefront only
Your store URL. Nothing connected.
Public catalogue, pages and reviews are enough to build the brain, audit the store, watch competitors, draft ads and draft the SEO plan. Works for any store on any platform.No performance numbers. The Monday brief says so rather than guessing.
AgentsBrain Builder
Competitor Watch
Store Fix (audit and drafted fixes)
Ad Creative (files delivered, nothing uploaded)
SEO and AI Search (audit and drafts) -
You paste the numbers
Last week's revenue, orders, ad spend, sessions when the Operator asks. Or an orders CSV.
The Weekly Brain Update produces a useful brief from as little as revenue and ad spend. Retention and Winback can segment from an orders export.Nothing is applied anywhere. Everything is still yours to paste.
AgentsWeekly Brain Update
Retention and Winback (from CSV) -
Connected, on your keys
Shopify custom distribution app with named scopes. Meta Ads. Klaviyo or Brevo. Your mailbox. Judge.me, Stamped or Reviews.io. Zernio. WhatsApp Cloud API. Instagram after Meta App Review.
Agents apply approved changes, upload paused ads, build disabled flows, send approved batches from your mailbox. Autonomous mode adds the schedule. The approval gate is identical.Theme code changes stay developer recommendations until Shopify grants the theme exemption. Instagram replies wait on Meta App Review, two to four weeks.
AgentsAll sixteen
Priced per agent. Bundled when cheaper.
Hosted means our infrastructure, our model tokens, the schedule, the approvals in Telegram and the Monday brief. Cancel any month.
Free for 14 days, no card
Competitor Watch and the Weekly Brain Update, hosted, for two weeks. Both are read only: nothing is written anywhere but our own storage. The first Monday digest has to earn the next thirteen days.
Core FourDefault set
$249₹14,999per month
Any four of the sixteen. The default set fixes the store, makes the ads, manages the account and finds the creators.
- Default set: Store Fix, Ad Creative, Ad Growth, Creator Outreach
- Swap any of the four for any other agent
- Cheaper than four single agents ($316)
- Approvals in Telegram, Monday brief, weekly review
The whole crew
$549₹34,999per month
All sixteen agents on the full weekly rhythm.
- All sixteen agents
- The complete seven day rhythm above
- Cheaper than eight single agents ($632)
- Approvals in Telegram, Monday brief, weekly review
Optional setup session90 minutes with a person. We build the brain with you, connect your tools on your keys and run the first agent together.$199Book the session
Founding price lock, first 100. The first 100 Hosted customers keep the price they signed up at for as long as they stay subscribed. The only urgency device on this site. No countdown.
USD and INR prices are the same product. Pick the currency you pay in.
Included volumes come from the agent contracts: Ad Creative makes up to 40 creatives a month, SEO drafts up to 10 programmatic pages a month, Creator Outreach sends at most 20 first emails a day.
Pick your agents. Watch the price.
Tick the agents you want. The total switches to a bundle the moment it is cheaper. Same numbers as above, nothing hidden.
Reference: what each count costs
- 1$79One agent
- 2$158Two single agents
- 3$237Three single agents
- 4$249Core Four
- 5$328Core Four plus one
- 6$407Core Four plus two
- 7$486Core Four plus three
- 8$549The whole crew
We do not compare this to salaries. Hiring costs depend on your city and your bar, and any number we put here would be invented. The comparison that is real is above: eight single agents would cost $632; the whole crew is $549.
What a performance marketer asks first.
What exactly gets a Telegram approval card?
Every live content change batch (Store Fix). Every campaign activation and every budget change (Ad Creative, Ad Growth). Every new ad set (Ad Growth). Every first email batch (Creator Outreach, UGC Seeding, Affiliate). Every flow enable (Lifecycle). Every campaign send and WhatsApp template (Retention, Post-Purchase). Every review reply and request (Reviews). Every page publish and discount code (SEO, Offer and Landing Page). Every scheduled post (Content Publishing). Every comment reply and DM (Social Engagement). The agent pages list each gate under Guardrails and approvals.
Does Autonomous mode skip approvals?
No. It changes when an agent runs, not what it may do without you. Ad Creative in Autonomous uploads ten fresh creatives paused every fortnight; activation still waits for your tap. Creator Outreach in Autonomous handles replies, but batch first emails still pause. The one exception is a per customer flag, off by default, that lets Creator Outreach accept a counter offer that is inside your budget band.
How is Meta handled?
Through the official Ads API. Every campaign, ad set and ad is created paused in a campaign named StoreCrew plus the date. The agent reads back effective_status and fails the upload if it is not paused. Ad Growth never moves a budget more than 20 percent in a day, never exceeds your daily cap, never creates a campaign without approval, and keeps new accounts in Advise for 14 days while the kill and scale rules are written and versioned in the brain.
Why Shopify GraphQL and a custom distribution app?
GraphQL only, no REST Admin paths. Custom distribution means the app installs on your store from a link without an App Store review. Scopes are named at install. One honest limit: Shopify now gates theme code changes behind an exemption we do not hold yet, so Store Fix writes product content, SEO fields, FAQ metafields and draft pages, and hands theme changes to your developer with the exact file.
What about Klaviyo, Judge.me, Instagram, WhatsApp?
Klaviyo with your private API key, flows created disabled. Brevo is supported too. Judge.me, Stamped and Reviews.io can send review requests through their APIs; Loox and Okendo are read only. WhatsApp through the Meta Cloud API with pre approved templates. Instagram through the Instagram Login API after Meta App Review, which takes two to four weeks, so Social Engagement drafts in Advise until then.
What is in the trial, really?
Competitor Watch: weekly snapshots of three to five competitors, a diff of products, prices over 5 percent, offers, collections, apps and theme, plus their new ads from the ad library, ranked by impact, maximum seven items, every item with a URL. Weekly Brain Update: the Monday brief built from what the brain knows, and it tells you plainly when it does not have your numbers. Both read only. No card.
How do I leave?
Plans are monthly. Email us. Your brain document is yours and we send it to you. Approvals that were pending expire. Anything the crew created lives in your own accounts: your paused campaign, your disabled flows, your draft pages, your codes.
Will it change my store without asking me?
No. Every write to your store, every message to a customer or creator, and every rupee or dollar of spend needs an approval you tapped. The approval is a one use token that expires. No record means no action. Silence is never consent. Store Fix, for example, asks once for the whole batch, applies it to product content only, and records every previous value so you can revert.
Can it spend my ad budget on its own?
No. Ad Creative uploads every campaign, ad set and ad paused, then reads back the status and fails the upload if anything is not paused. Activating a campaign and setting a budget are separate approval gated actions. Ad Growth keeps new accounts in Advise for the first 14 days, never moves a budget more than 20 percent in a day, never breaches your daily cap, and pausing a loser is the only thing it does without asking, because pausing costs nothing.
What access does it need?
Level 1 needs nothing but your store URL: public catalogue, pages and reviews are enough for the brain and for Competitor Watch, Store Fix audits and Ad Creative drafts. Level 3 means you connect tools on your own keys: a Shopify custom distribution app with named scopes, Meta Ads, your email tool, your mailbox, your review app. Each agent page lists exactly which tools it uses and in which mode.
Does it email my customers or creators by itself?
First emails always pause for your approval, in every mode, including Autonomous. Creator and affiliate outreach sends only from your own connected mailbox, never from our domain, at most 20 first emails a day, and never a fourth message to a creator who stayed silent. There are no cold Instagram DMs. Ever. A private reply goes only to someone who commented on your post, once, within seven days.
Is this just ChatGPT with a long prompt?
The Brain + Crew tier does run inside your own ChatGPT or Claude, on your tokens, and we say so. What you are buying is the brain: a versioned document with six layers that every agent reads before it works and writes back to after. Plus sixteen agent contracts with workflows, quality gates and guardrails you can read on this site. Hosted adds the tools, the approvals over Telegram and the weekly schedule.
What is the brain, concretely?
A living, versioned document about one store. Six layers: identity, customers, products, competitors, performance, learnings. The Brain Builder writes version 1 from your storefront, your public reviews and six answers from you, and marks anything it inferred. Only the Weekly Brain Update may write performance. Learnings are append only, so nothing gets quietly rewritten.
Will it make up claims about my products?
It cannot. The brain holds claims_allowed, built only from text found on your own storefront, and claims_forbidden, built from your category's regulated claims. Every drafted line is checked against both. Anything that fails is rewritten or removed and listed as needs your confirmation. Ad Creative ships a claims ledger with every batch: every claim, its source, allowed or rewritten.
I am not on Shopify. Can I use it?
Yes, at level 1. Storefront only mode works for any store because the public catalogue and pages are enough for the brain, Competitor Watch, the Store Fix audit and Ad Creative drafts. Act mode writes to a store through Shopify GraphQL first. Other platforms come later and we will not pretend otherwise.
How do I undo something?
Store content changes ship with a changes file that holds every previous value, so a rollback is re applying the old value. Ads are created paused in a campaign named StoreCrew with the date, so you can see and pause everything the crew made. Flows are created disabled. Pages are created as drafts. You can pause any agent, and a hosted plan is monthly.
Where does my data live and who can see it?
Under your customer id. Every table carries it and every query filters by it. Competitor Watch never logs in, never submits a form, never adds to cart. Google review text is never stored. Secrets live in a secrets store, not in code or logs, and tokens are redacted from every log line.
Why do you not show ROAS or revenue numbers?
Because we would have to invent them. The site shows the mechanism and the price. Inside the product, the Weekly Brain Update reports revenue divided by spend and always labels it blended, never repeats platform reported ROAS, and writes likely because with evidence or no clear cause. Never a bare causal claim.
What is the founding price lock?
The first 100 Hosted customers keep the price they signed up at for as long as they stay subscribed. That is the only urgency device on this site. There is no countdown.
It runs. It sends. You read.
Put the crew on your store.
Start with the two read only agents free for 14 days. Read the Monday digest. Decide with evidence.
Founding price lock for the first 100 Hosted customers. No countdown.