OrgGen AI

Official registry AI and agents

Human-agent interface for projects: you prioritize and decide, your AI agents execute. Over MCP.

Install

Add this to your client's MCP configuration (Claude Desktop: claude_desktop_config.json; Cursor: .cursor/mcp.json; VS Code and others use the same "mcpServers" shape).

Remote (streamable-http)
Claude Code
claude mcp add --transport http orggen-ai 'https://www.orggen.ai/api/mcp' --header 'Authorization: <Bearer token. Generate an API key from https://www.orggen.ai/settings — free with an account (rate-limited to 500 calls/day); higher rate limits on paid plans.>' --header 'X-Project-Id: <Optional. UUID of the project to operate on for this request; it overrides the key'\''s stored active project. Tools also accept expectedProjectId on each call.>'
JSON configuration (Claude Desktop, Cursor, VS Code)
{
    "mcpServers": {
        "orggen-ai": {
            "type": "http",
            "url": "https://www.orggen.ai/api/mcp",
            "headers": {
                "Authorization": "<Bearer token. Generate an API key from https://www.orggen.ai/settings — free with an account (rate-limited to 500 calls/day); higher rate limits on paid plans.>",
                "X-Project-Id": "<Optional. UUID of the project to operate on for this request; it overrides the key's stored active project. Tools also accept expectedProjectId on each call.>"
            }
        }
    }
}

Documentation

From the project's README on GitHub, refreshed weekly. Links open on GitHub.

Contents
  1. OrgGen AI — MCP server
  2. What agents can do with it
  3. Connect
  4. Use
  5. About

OrgGen AI — MCP server

OrgGen is the human-agent interface for your projects. You set priorities and settle decisions; your AI agents work from the same board without waiting on you: they pull the next ready task, can raise a decision with a recommendation instead of stopping to ask, and keep the project diagrams (architecture, data model, process, user journeys, org charts, cap tables) up to date. Standard HTTP MCP server. The free plan includes MCP with an account.

This repository is documentation only. OrgGen is a hosted service: the MCP server runs at the endpoint below and its source code is not published here. This page describes what the server does, how to connect to it and how to use it.

What agents can do with it

An agent connected to OrgGen reads and edits the projects of the account that authorized it:

  • Diagrams on a shared canvas: organization charts, cap tables, data models (entities, fields, relations), architecture, process flows and user journeys.
  • Tasks and decisions: create, claim and update tasks; raise a decision with options and a recommendation, and read the answer later.
  • Documents: rich documents attached to blocks of a project.
  • Project knowledge: the server explains how to use it. It sends instructions when a session starts, and a get_playbook tool returns the rules for each kind of diagram.

The list of tools is served by the server itself (tools/list), so it is always current. Tools can create, change and delete content in the projects of the account that authorizes the connection.

Connect

With OAuth 2.0 (no API key)

The server supports OAuth 2.0 (authorization code with PKCE S256). You sign in to OrgGen, check which application is asking for access, and allow or deny it.

OAuth has been verified with Claude (custom connector in claude.ai) on 2026-10-22. We have not tested it with other clients, and the server currently authorizes only the clients it has been proven with. Other clients can connect with an API key.

With an API key
  1. Create a free account at https://www.orggen.ai.
  2. Generate an API key in your account settings.
  3. Send it as a bearer token with every request:
Authorization: Bearer YOUR_API_KEY

Many MCP clients accept a URL plus custom headers. The exact configuration file depends on the client; the values are:

{
  "url": "https://www.orggen.ai/api/mcp",
  "headers": { "Authorization": "Bearer YOUR_API_KEY" }
}

The free plan includes MCP access with an account, rate-limited to 500 calls a day; paid plans have higher limits.

Use

  • One API key per agent. A key remembers an active project; two agents sharing a key can overwrite each other's target.
  • Target the project on every write. Pass expectedProjectId (the project's UUID) on any call that modifies data. Optionally send an X-Project-Id header to choose the project for a request.
  • Read the checks field of a write response: it lists business problems that your call has just introduced (for example, shares above 100% in a cap table). Silence means nothing to report.
  • Start with get_schema_overview to see what a project already contains before creating anything.

About

OrgGen AI is a hosted service operated by its publisher; see the terms of service linked above. This repository contains no service source code and carries no license for it.

Frequently asked questions

What is OrgGen AI?

Human-agent interface for projects: you prioritize and decide, your AI agents execute. Over MCP.

How do I install OrgGen AI?

It runs as a remote server. In Claude Code: claude mcp add --transport http orggen-ai 'https://www.orggen.ai/api/mcp' --header 'Authorization: <Bearer token. Generate an API key from https://www.orggen.ai/settings — free with an account (rate-limited to 500 calls/day); higher rate limits on paid plans.>' --header 'X-Project-Id: <Optional. UUID of the project to operate on for this request; it overrides the key'\''s stored active project. Tools also accept expectedProjectId on each call.>'. For other clients, paste the JSON configuration from this page into their MCP settings.

Does OrgGen AI need an API key?

Yes. It expects Authorization, X-Project-Id, which you set in its configuration.

Is OrgGen AI open source?

Its code is public on GitHub, but the repository does not declare a standard license.

Who maintains OrgGen AI?

It is published by OrgGen-AI. AI.tn lists it but does not maintain or endorse it.

From the official MCP Registry. Listing is not an endorsement: review a server's code and permissions before giving it access to your data.