Is there a public API?

Yes. InstinctGTM's REST API lives at /api/v1 and is described at /api/v1/openapi.json. Every route runs a tool from the same registry as the MCP server.

Can the API send or publish?

No route sends or publishes on its own. Approving a draft is its own call, open only to a key granted that permission, and the send runs from the queue after it.

Can my backend send website events?

Yes. A key holding only website events can post sign-ups and payments to /api/v1/events for website analytics. It reads nothing and drafts nothing.

The same tools, over plain HTTP

For a script or a service that cannot run a browser sign-in. Scoped keys, one OpenAPI document, and the same approvals as the screen.

Get my free GTM plan See pricing

OpenAPI document
Scoped keys, shown once
Rate limits in every response

The short version

InstinctGTM's REST API at /api/v1 runs the same tools as the MCP server through the same permission checks, described in one OpenAPI document generated from the registry the server runs on. You issue a key with only the scopes it needs; the secret is shown once and stored hashed, and rate limits come back in headers on every response.

Read against the registry on 2026-10-13

Set it up

1
Issue a key
Open Settings, then Agent access, and issue a key with only the scopes it needs, for only the ventures it should reach.
2
Read the document
Fetch /api/v1/openapi.json. It lists every route below, generated from the same registry the server runs.
3
Call it
Send the key as a bearer token. Rate limits come back in RateLimit headers, so you never have to find them by hitting them.

Open the OpenAPI document

Every route

30 routes under /api/v1, each running one tool.

GET	/api/v1/ventures	The ventures this credential can reach.
GET	/api/v1/brief	Everything waiting, in one call.
GET	/api/v1/approvals	Drafts a human has yet to allow.
GET	/api/v1/followups	People owed a next touch.
GET	/api/v1/companies	Search watched companies.
POST	/api/v1/companies	Add a company to the watchlist.
GET	/api/v1/companies/{company_id}	One company, with its dated timeline.
GET	/api/v1/companies/{company_id}/explain	Why the system concluded what it did.
POST	/api/v1/companies/{domain}/read	Ask for a company to be re-read.
GET	/api/v1/people	Find people in the CRM.
POST	/api/v1/people	Add a person.
GET	/api/v1/people/{person_id}	One person, with their history and intel.
POST	/api/v1/people/{person_id}/touchpoints	Record that something happened.
POST	/api/v1/people/{person_id}/stage	Move a relationship to a new stage.
POST	/api/v1/people/{person_id}/snooze	Push a follow-up out, or clear it.
POST	/api/v1/people/{person_id}/followup	Draft the next message in a sequence.
GET	/api/v1/content	Find content by channel and state.
POST	/api/v1/content	Turn one idea into a post per channel.
POST	/api/v1/content/{artifact_id}/approve	Approve one piece. Does not publish.
GET	/api/v1/posts	The posting calendar: drafts, scheduled, failed and published posts.
POST	/api/v1/posts	Save a post as a draft for the networks named.
POST	/api/v1/posts/{group_id}/schedule	Schedule a drafted post. At its time it goes out.
POST	/api/v1/content/{artifact_id}/published	Record that an approved piece went out.
POST	/api/v1/ideas	Queue an idea without drafting it.
GET	/api/v1/mentions	Mentions and captures worth answering.
POST	/api/v1/mentions	Hand over a thread worth answering.
POST	/api/v1/mentions/{mention_id}/reply	Draft a reply to somebody who engaged.
POST	/api/v1/outreach/{action_id}/approve	Approve one outreach draft. This sends.
GET	/api/v1/performance	What each channel produced against what it cost.
GET	/api/v1/learning	Aggregate patterns. Counts only, no identities.

Keys, by what they can do

read: Reads only.
capture: Reads, and queues watchlist entries, people, notes, captures and ideas.
draft: Reads, queues and drafts into your approvals. Cannot approve.
full: Everything an outside caller can hold, including approving. Still cannot send without an approval.
events: Only website events from your own backend.

What no credential can do

send or publish
Sending and publishing go out under the founder’s name. They run from the queue after an approval row exists, and creating that row is a separate, explicit call. An agent that can approve and send in one step is an agent that can publish without anybody reading it.
raise a limit, a cap or an autonomy stage
Deliverability caps, per-domain limits and the autonomy ladder are the blast-radius controls. A caller that can widen its own limits has no limits.
read across ventures without a grant
A venture is the isolation context. A key names its ventures and every read is filtered by that list, not by what the caller asks for.
bulk export contacts or message bodies
Page-at-a-time reads serve every legitimate agent workflow. An unbounded dump of people and message bodies is the shape of an exfiltration, and it is not needed for anything the product does.
write to the learning boundary
Outcomes are recorded by the pipeline that observed them. A caller that can write learning rows can teach the allocator anything it likes, and the allocator moves real budget.
reach billing, keys, or another member’s session
Account administration is a human surface. Nothing about an agentic workflow requires it, and everything about it is worth a password prompt.

Questions people ask