@Atlas audits the catalog the day the store is connected and ships a CRO plan, instead of handing back a list of things to go read. Audit to plan, same session.
BoostEcom DemoDemo
Founder
Browse documentation
Documentation
The reference for BoostEcom: how the product works, and how to wire an agent to it. Every figure on this page is derived from the codebase.
The surface
What there is to document
- API route handlers
- 388
- Registered MCP tools
- 15
- MCP scopes
- 9
- Webhook endpoints
- 7
Using BoostEcom
What the platform does, how the AI team works, and what each plan includes.
Getting started
- What BoostEcom isA Commerce Operating System where the AI operates the business and you own it. What that means concretely, and what it does not mean.
- QuickstartConnect a Shopify store and get your first audit back. Five steps, no configuration files, no code.
- Autonomy levelsThe four levels that decide how much the AI may do without asking you. What changes at each one, and what never changes.
- Connect your storeTwo ways in: connect a Shopify store you already run, or have one provisioned for you. What each grants, and what survives an ownership transfer.
- Organizations & rolesThe tenant boundary, the five roles and what each one can do, per-member overrides, and how an invitation actually works.
How it works
- The AI teamAtlas orchestrates five specialists. Who owns what, how they coordinate, and how to address one directly.
- SystemsThe six productised capabilities that run against a connected store: what each one reads, and what it gives back.
- Chat & channelsWhere you can talk to the AI team, what a conversation holds, what happens when a connection drops mid-answer, and what each channel reserves before it streams.
- MemoryWhat the AI remembers between conversations, at which scope, how a fact decays, and how to correct or delete one.
Plans & credits
- Plans & limitsWhat's included in each BoostEcom plan, the credit ledger model, message and store caps, and what triggers an overage.
- Credits & meteringHow the credit ledger works, the two guardrails that stop an overrun, and why your renewal date is not the first of the month.
- BillingWhat Stripe charges and when, why your renewal is your anniversary rather than the 1st, how tax is handled, and where to change or cancel a plan.
Systems
- Tracking ScannerAudits a storefront's tracking stack (tags, pixels, consent mode, datalayer) and returns a graded, actionable report. The one System with a public entry point.
- SEO AuditorTechnical and semantic SEO audit of a connected storefront, run against real Shopify data rather than a crawl-only guess.
- AEO OptimizerAnswer Engine Optimization: the three conditions for an LLM to cite your brand, audited weekly, plus a monthly reading of the citations you actually get.
- Conversion AnalyzerCRO diagnosis built from your own Shopify orders, checkout settings and payment methods: a prioritised action plan, not a benchmark table.
- AOV OptimizerBundles ranked by measured lift from products your customers actually bought together, and a free-shipping threshold derived from your own order distribution.
- Web VitalsField performance from three fused sources (your real visitors, Chrome's CrUX dataset and PageSpeed Insights), with the most trustworthy one picked per metric.
Trust
- SecurityHow you sign in, how long a session lasts, where the tenant boundary is enforced, how tokens are stored, and what is written to the audit log.
- Data & privacyWhat is kept and for how long, how to export or erase it, and how the three mandatory Shopify erasure webhooks are actually drained.
- Status & incidentsNine probed services, a rolled-up 90-day history, how incidents are published, and the three ways to be told about one without watching a page.
Buying & selling stores
Listings, offers, the deal pipeline, escrow payouts, KYC and disputes.
Buying
- Marketplace overviewTwelve public categories, what each one supports, how a listing gets published, and what it costs.
- Buying on the marketplaceCheckout, taxes, what a subscription listing bills each month, where your orders live, and the fourteen days you have to dispute.
- Acquiring a storeThe deal pipeline from offer to close: who is allowed to advance each stage, what proof each transition demands, and why the buyer closes.
- Deal documents & KYCNDA, LOI and APA on a deal thread, how signatures are recorded, and the single KYC gate in the pipeline: where it sits, and why there.
- DisputesFourteen days to raise one, seven before it escalates to a mediator, the seven reasons, and the four ways one ends.
Selling
Building on BoostEcom
The MCP server, the credential families, the webhooks and the Intelligence API.
API
- API overviewThe shape of the BoostEcom API surface: what is public, what is authenticated, and which door to take for which job.
- AuthenticationThe credential families BoostEcom issues, what each one opens, and why none substitutes for another.
- Errors & rate limitsThe status codes the API returns, what each one means in this platform, and what to do when you get a 402 or a 429.
- VersioningThere is no API version header. What the two MCP paths mean, what is stable, and what happens when something is eventually deprecated.
MCP server
Agents & crawlers
Where do I start?Docs
Depends what you are doing. Operating a store: the Quickstart. Wiring an agent: the API overview, then Authentication. Both trunks are listed in full on this page, so there is no hidden section and no separate docs site.
Which token for which use?Docs
OAuth 2.1 with PKCE for an MCP client that can negotiate, a bst_mcp_ bearer for a client configured by a file, and a read-only bei_ key for the Intelligence API. A fourth prefix, bst_, is admin-internal and not issued to customers.
How is versioning handled?Docs
There is no version header. The MCP endpoint has an unversioned path kept for already-installed clients and a /v1/ path for new ones, both backed by the same handlers. Nothing has been deprecated, so no retention window is published.
Can an AI agent read this documentation?Docs
Yes, and that is deliberate. The site publishes a curated markdown index at /llms.txt, its structured variant at /llms.json, and robots.txt explicitly allows GPTBot, ClaudeBot, anthropic-ai, Perplexity and Google-Extended. Every docs page renders server-side, so fetching the URL returns the full text.
Something is missing or wrong. What do I do?Docs
Open a forum thread. The repository is private, so there is no pull request to send. This page previously said otherwise, which was simply false. The forum is read, and a reported gap is the fastest route to a fix.
One operator, the full stack: Maya runs the email calendar, Marco curates collections, Otis watches fulfilment. One team, one thread.
BoostEcom DemoDemo
Head of Growth
The work gets done rather than suggested. Voice mode with @Atlas turns the morning check-in into a conversation. It operates.
BoostEcom DemoDemo
Store owner
@Maya rewrites product descriptions, queues them for review, pushes the approved batch live, then runs the campaign around them. End-to-end, with approval gates.
BoostEcom DemoDemo
E-commerce lead
Every action the AI Team takes is graded against real Shopify Analytics after the observation window closes. Outcome attribution, not vibes.
BoostEcom DemoDemo
Marketing lead
OAuth scopes, audit log, store-scoped credits, no shadow data. The agents work inside the admin, not around it.
BoostEcom DemoDemo
Engineer
Permission mode decides what ships on its own and what waits for a human. Autonomy you dial in.
BoostEcom DemoDemo
Operator
Multi-store from one console, each with its own connectors, memory and budget. Per-store context.
BoostEcom DemoDemo
Agency owner
@Atlas audits the catalog the day the store is connected and ships a CRO plan, instead of handing back a list of things to go read. Audit to plan, same session.
BoostEcom DemoDemo
Founder
One operator, the full stack: Maya runs the email calendar, Marco curates collections, Otis watches fulfilment. One team, one thread.
BoostEcom DemoDemo
Head of Growth
The work gets done rather than suggested. Voice mode with @Atlas turns the morning check-in into a conversation. It operates.
BoostEcom DemoDemo
Store owner
@Maya rewrites product descriptions, queues them for review, pushes the approved batch live, then runs the campaign around them. End-to-end, with approval gates.
BoostEcom DemoDemo
E-commerce lead
Every action the AI Team takes is graded against real Shopify Analytics after the observation window closes. Outcome attribution, not vibes.
BoostEcom DemoDemo
Marketing lead
OAuth scopes, audit log, store-scoped credits, no shadow data. The agents work inside the admin, not around it.
BoostEcom DemoDemo
Engineer
Permission mode decides what ships on its own and what waits for a human. Autonomy you dial in.
BoostEcom DemoDemo
Operator
Multi-store from one console, each with its own connectors, memory and budget. Per-store context.
BoostEcom DemoDemo
Agency owner
Something missing or wrong?
Tell us in the forum: it is the fastest way to get a doc fixed.