Skip to content

Connect your AI client ​

Plug Claude, ChatGPT, Codex, or any MCP-aware agent into your Genudo workspace.

Everything below points at one endpoint and authenticates the same way — your client opens a browser sign-in and you approve access to your Genudo account. There is no token to create or paste.

Server URL ​

https://api.genudo.ai/mcp

Every command and config block on this page uses that URL, and it always points at the environment you are reading these docs on — copy it from here rather than typing it from memory.

Transport is Streamable HTTP (MCP spec 2025-03-26). Auth is OAuth — the first request returns 401 with discovery headers, your client registers itself, and the browser login follows automatically. Approve the mcp:use scope and you're connected.

Options at a glance ​

OptionRatingRuns onWhat you get
Plugin ⭐ Recommended★★★★★Claude · ChatGPT · Codex · CoworkConnector + skills + agents
One command★★★★★Claude Code · Codex CLI · Gemini CLI · VS Code32 tools, one line in your terminal
In the app★★★★☆Claude Desktop · claude.ai · ChatGPT · Cursor · Windsurf32 tools, no terminal
Direct HTTP★★★☆☆Custom agents & SDKsRaw JSON-RPC with a Bearer token

The plugin is the full experience; every other option is the connector on its own — the same 32 tools, none of the skills or agents.


Plugin ​

Recommended · The connector plus the skills and agents that know how to build, edit, and audit your pipelines

Pick the ecosystem you work in. Both install from a marketplace repository and prompt for no configuration at all — you approve the browser sign-in on first use.

Claude ​

Runs in Claude Code, Claude Desktop, claude.ai, and Cowork. Includes 32 tools · 22 skills · 6 agents.

Marketplace: genudo-ai/claude-plugin

In the app

  1. Open the plugin marketplaces and add one from a repository Settings → Plugins → Add → Add marketplace → Add from a repository
  2. Paste the marketplace above, then sync it genudo-ai/claude-plugin → Sync
  3. Choose the full plugin — "Genudo - AI Workforce" — and install it Personal → claude-plugin → Genudo - AI Workforce
  4. Open its Connectors tab and install the genudo connector Connectors → genudo → Install
  5. Approve the sign-in in your browser — done

From Claude Code

Type both in a Claude Code session:

text
/plugin marketplace add genudo-ai/claude-plugin
/plugin install genudo@genudo-ai

ChatGPT ​

Runs in the ChatGPT desktop app and Codex. Includes 32 tools · 28 skills.

Marketplace: genudo-ai/chatgpt-plugin

In the app

  1. Open the plugin marketplaces and add one Plugins → Add → Add a marketplace
  2. Paste the marketplace above as the Source, then add it Source → genudo-ai/chatgpt-plugin → Add marketplace
  3. Open "Genudo - AI Workforce" from your personal plugins Personal → Genudo Plugins (ChatGPT) → Genudo - AI Workforce
  4. Hit Try now, then approve the sign-in in your browser — done

From the Codex CLI

Run in your terminal, then sign in:

bash
codex plugin marketplace add genudo-ai/chatgpt-plugin
codex mcp login genudo

What the plugin adds

Skills and agents are what you get on top of the raw tools — see Skills & agents for the full list and paste-ready prompts. Every write is previewed as a diff and confirmed before it touches your account.


One command ​

No install, no Node.js, no token paste — one line per client, then approve the browser sign-in.

bash
claude mcp add --transport http genudo https://api.genudo.ai/mcp
bash
codex mcp add genudo --url https://api.genudo.ai/mcp
codex mcp login genudo
bash
gemini mcp add --transport http genudo https://api.genudo.ai/mcp
bash
code --add-mcp "{\"name\":\"genudo\",\"type\":\"http\",\"url\":\"https://api.genudo.ai/mcp\"}"

Verify the connection from inside the client — /mcp in Claude Code and Codex, /mcp list in Gemini CLI, or the MCP panel in VS Code. You should see genudo connected with 32 tools.


In the app ​

No terminal. Copy the server URL, open the screen below, paste, and sign in.

Claude Desktop · claude.ai ​

Claude → Settings → Connectors → Add custom connector. Name it Genudo and paste the URL.

ChatGPT ​

Settings → MCPs → Add server. Choose Type: Streamable HTTP, name it Genudo, and paste the URL.

Codex ​

Add it to your Codex config:

json
{
  "mcpServers": {
    "genudo": { "url": "https://api.genudo.ai/mcp", "auth": "oauth" }
  }
}

Then run codex mcp login genudo to sign in.

Cursor · Windsurf · Zed · any other client ​

Any client that speaks Streamable HTTP uses the same block — paste it into that client's MCP config (~/.cursor/mcp.json for Cursor globally, or .cursor/mcp.json in a project):

json
{
  "mcpServers": {
    "genudo": { "type": "http", "url": "https://api.genudo.ai/mcp" }
  }
}

Restart the client after saving. On first use it opens the browser sign-in.


Direct HTTP (advanced) ​

For custom agents and SDKs that call the endpoint directly. This is the one path that uses a Personal Access Token instead of the browser sign-in — create one with the mcp:use scope from the Authentication guide.

Claude Code (with an explicit Bearer token):

bash
claude mcp add --transport http genudo \
  https://api.genudo.ai/mcp \
  --header "Authorization: Bearer YOUR_TOKEN_HERE"

TypeScript:

typescript
import { Client } from '@modelcontextprotocol/sdk/client/index.js';
import { StreamableHTTPClientTransport } from '@modelcontextprotocol/sdk/client/streamableHttp.js';

const transport = new StreamableHTTPClientTransport(
  new URL('https://api.genudo.ai/mcp'),
  { headers: { Authorization: 'Bearer YOUR_TOKEN_HERE' } }
);

const client = new Client({ name: 'my-agent', version: '1.0.0' }, { capabilities: {} });
await client.connect(transport);

Python:

python
from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client

async with streamablehttp_client(
    'https://api.genudo.ai/mcp',
    headers={'Authorization': 'Bearer YOUR_TOKEN_HERE'}
) as (read, write, _):
    async with ClientSession(read, write) as session:
        await session.initialize()
        tools = await session.list_tools()

curl:

bash
curl -X POST https://api.genudo.ai/mcp \
  -H "Authorization: Bearer YOUR_TOKEN_HERE" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Troubleshooting ​

SymptomCauseFix
Browser sign-in never opensClient doesn't support OAuth discoveryUse Direct HTTP with a Bearer token instead
401 UnauthorizedSign-in was cancelled, or the token is invalid/expiredReconnect the server, or re-issue the token
403 Missing required scope: mcp:useToken missing mcp:use scopeCreate a new token with mcp:use selected
Individual tool calls fail with 403Token missing a tool-level scopeRe-issue the token with the required scope added
Tools not showing after restartStale desktop-app cacheFully quit and relaunch (Cmd/Ctrl+Q, not just close the window)
Skills or agents missingThe connector is installed, not the pluginInstall the Plugin — the connector alone carries tools only