Do MCP servers need to be publicly accessible?

Yes. The MCP server must be reachable over HTTP from GAIA's servers. Local development servers are not directly supported unless exposed via a tunnel like ngrok.

How does GAIA handle authentication for MCP servers?

Servers requiring bearer tokens accept a token added during setup. OAuth-based servers are handled via spec discovery — GAIA follows the server's auth flow automatically.

Are MCP tools available to subagents?

Yes. All connected MCP tools are shared across the main agent and every specialized subagent.

All features

Integrations

Connect any AI tool, not just GAIA's list.

GAIA supports the Model Context Protocol — connect any MCP-compatible server and its tools become immediately available to every GAIA agent.

Server URL

https://mcp.github-tools.io/v1

Connect

Integrations

Key capabilities

What makes MCP Support powerful.

Any MCP server

Point GAIA at any HTTP MCP endpoint and its tools are auto-discovered and indexed.

Auth-aware

MCP servers requiring OAuth are handled automatically via spec discovery.

Extends subagents

MCP tools are available to the main agent and specialized subagents alike.

How it works

Three steps to get started.

Set up in minutes. Works automatically from there.

Paste the MCP server URL

Enter the HTTP endpoint of any MCP-compatible server in the MCP settings panel.

Tools are auto-discovered and indexed

GAIA reads the server's manifest, discovers all available tools, and adds them to its toolkit.

Use tools in chat or workflows

All MCP tools are available immediately in every conversation and workflow, alongside GAIA's native tools.

Use cases

How teams use this.

Real workflows, real outcomes.

Internal search tool via MCP

An engineering team wraps their internal documentation search as an MCP server and connects it to GAIA, making internal docs searchable from any conversation.

Proprietary database query tool

A data team exposes a read-only SQL query endpoint as an MCP server, giving GAIA the ability to fetch live production metrics directly in workflows.

Third-party AI tool integration

A developer connects a third-party vector search service that supports MCP, making semantic search across their knowledge base available to every GAIA workflow.

FAQ

Frequently asked questions.

Everything you need to know.

Model Context Protocol is an open standard for connecting AI agents to external tools. Any service that implements MCP can expose its capabilities to GAIA.

Features that work well together.

Combine these with MCP Support for more power.

50+ Integrations

Connect Gmail, Slack, GitHub, Notion, and 47 more

Custom Integrations

Build, publish, and share your own integrations

Specialized Agents

37 purpose-built agents — one for every integration