Docs

Integrations & MCP

Connect tools, data, and models to your agents through the Model Context Protocol.

Agents are only useful when they can act. Padiso connects them to the outside world through the Model Context Protocol (MCP).

How connections work

Connect a tool, API, database, or model once at the workspace level. Every agent in the workspace can then use it, with no per-agent wiring.

Two directions

Padiso as an MCP server. A workspace-scoped MCP endpoint exposes your agents, tasks, and projects to any MCP client, including Claude.ai, the CLI, and your own apps.
Padiso as an MCP client. Your agents use any MCP server, API, database, or internal service you connect.

What you can connect

SaaS tools: Slack, GitHub, Linear, Notion, Salesforce, Google Drive, and more.
Your own internal APIs and databases, exposed as MCP tools.
Models and runtimes (see Runtimes & models).

Browse everything in the integrations directory, or read more on the MCP page.

Scoped and governed

Every connection is scoped to the capabilities you allow, sensitive actions are gated behind approvals, and every call is written to an audit trail. See Governance.

Concepts Runtimes & models