MCP and API · ontology/

Every app's ontology as an API and MCP tools, checked per caller.

Deploy an ontology and it is served four ways: a REST API, an OpenAPI reference, a typed SDK for your apps and an MCP server for agents. People connect their agents with OAuth and the agent acts as them, with tools only for what their roles allow.

cloud.segmentstream.com
The console's access view: every read and action of the ontology in rows, the roles in columns, with a dot where the role holds the permission
Who may do what, computed from the ontology. The same table applies to people, API keys and agents.

An MCP server per app

Every app in a workspace serves its ontology at /mcp. Add the address to Claude, ChatGPT, Claude Code or any MCP client that supports OAuth.

Agents sign in as the person

The member signs in on the app's own page and chooses: everything their roles allow, read only, or deny. The agent then acts as them, and history names them.

Tools per role

tools/list holds only the tools the caller's roles allow, and every call is checked again. A role without deal:moveStage never sees deal_move_stage.

REST, OpenAPI and a typed SDK

The same model is a REST API at /api/ontology with an OpenAPI reference, and a typed client for your apps. One code path serves all of them, so a rule enforced in one is enforced in all.

API keys for unattended work

Jobs that run with nobody signed in use API keys. A key holds exactly the roles it was created with.

Errors an agent can act on

A refused call returns the action's own message, so the agent can fix its input and try again. Clients ask before running an action that changes data.

In code

terminal
# A member connects their agent, then signs in on the app's page.$ claude mcp add --transport http crm https://crm.example.com/mcp# The tools it gets depend on the member's roles, for example:describe  me  aggregatedeal_find  deal_get  deal_history  deal_move_stagelead_find  lead_get  lead_history  lead_claim  lead_qualify# Unattended work uses an API key with one role.$ segmentstream ontology mcp --create-key nightly-report --role sales
MCP and API reference →

Use it, build on it, or have it built

Use it

Copy your app's /mcp address into Claude or ChatGPT and sign in.

Build it with your agent

Add an action or a role in ontology/ and deploy; the API and the tool list follow.

Have our engineers build it

Our engineers connect your team's agents and set up the keys your other systems need.

Replaces Hand-built MCP servers, Database credentials handed to agents and Vendor APIs with dozens of tools to choose from.

Works with

Solutions built with it

Read more

Enterprise pilot

Start a pilot.

Want mcp and api set up for you? Our forward-deployed engineers build it with you in your workspace, as code you own.

  • One solution, on your data, in your own workspace
  • First version live in two weeks
  • Parity proven before anything is switched off
  • Fixed scope and price, credited to your contract if you continue

Not ready for a pilot? Send the form anyway with your question. An engineer will reply, not a sales rep.

How the pilot works

Goes to our founder’s inbox. No sales queue, no newsletter.

Your company’s software, built by agents, owned by you.

$ segmentstream workspace create acme --region eu --subdomain acme

▸ Created workspace acme · database, warehouse and a private git repository

$ segmentstream workspace WORKSPACE-ID clone acme && cd acme

▸ your agent writes the five folders and runs the checks

$ git add -A && git commit -m "First version"

$ segmentstream deploy

▸ app, ontology API and MCP tools live for your invited members

Start a pilot

Our engineers build the first version on your data in two weeks, at a fixed price.

Or join the waitlist for self-serve:

We’ll email you product updates, tips and onboarding help. Unsubscribe anytime.

Self-serve opens to the waitlist first.