ConnectKit

The local connector runtime
for AI agents.

ConnectKit is an alpha framework for turning user SaaS accounts into agent tools with YAML connector specs, a local OAuth router, encrypted per-user credentials, and CLI or MCP adapters.

Hybrid DB SQLite + FTS5 Rows · Full-text ChromaDB Vectors · Semantic DuckDB Analytics · OLAP Graph Traversal · Rank

Account-connected tools without outsourcing the runtime.

hub

YAML-defined connectors

Add services with declarative specs instead of writing custom Python for every SaaS API.

key

Universal OAuth router

One login and callback flow handles OAuth providers with encrypted state and token exchange.

encrypted

Encrypted local vault

Per-user SQLite credential vaults store tokens with Fernet encryption and no external service.

terminal

CLI adapters

Wrap existing tools like gws or gh and inject user credentials into subprocess environments.

extension

MCP adapters

Expose connected SaaS accounts to MCP servers with per-user credential injection.

smart_toy

Agent-native meta-tools

Agents can list connectors, connect accounts, disconnect, and inspect health as normal tools.

Specs, vault, runtime, adapters, and OAuth in one local package.

ConnectKit loads connector specs, checks each user's encrypted vault, discovers tools for connected services, and returns agent-ready tool definitions through the same pattern as the MCP bridge.

ConnectorRuntime

Loads YAML specs, checks connection state, selects adapters, and discovers tools.

CredentialVault

Encrypted SQLite storage for per-user OAuth tokens, API keys, and connection metadata.

OAuth router

FastAPI routes for login, callback, status, state validation, and token exchange.

CLI and MCP adapters

Turn connected services into executable, agent-readable tools.

The value prop: own the bridge between users and tools.

check_circleNo hosted connector platform required; the OAuth router and vault run locally.

check_circleConnector catalog includes setup fields, status, icons, categories, and descriptions for UI use.

check_circleTool descriptions are optimized for LLM discovery rather than only human API docs.

check_circleBest fit for developer-controlled agents where users can install CLIs or run MCP servers.

Nango-style auth meets agent-native local tools.

Hosted platforms like Composio, Pipedream, Zapier, and Arcade offer huge integration catalogs. ConnectKit is for teams that want the connector runtime, vault, and specs under their own control.

Versus Nango

Nango is strong product integration infrastructure. ConnectKit borrows the provider-catalog idea but focuses on local agent tools, encrypted local vaults, and CLI/MCP execution.

Versus Composio and Arcade

Those are polished agent-tool platforms with managed auth and large catalogs. ConnectKit trades hosted convenience for local control and inspectable connector specs.

Versus Zapier and Pipedream

Zapier and Pipedream win on breadth and no-code automation. ConnectKit is code-first infrastructure for embedding account-connected tools into your own agent runtime.

Versus WorkOS or Clerk

WorkOS and Clerk authenticate your app's users. ConnectKit handles downstream SaaS authorization: the user's Gmail, GitHub, Slack, Drive, and tool credentials.

Use ConnectKit when you want owned infrastructure, not a hosted marketplace.

radio_button_checkedUse it for local or private agents that need per-user SaaS credentials and agent-readable tools.

radio_button_checkedTreat the 544+ specs as a catalog baseline, not a promise that every connector is production-tested today.

radio_button_checkedHTTP adapters and full token lifecycle management should not be overclaimed until implemented end-to-end.

radio_button_checkedThe clearest CTA is not 'connect to every app instantly'; it is 'own the connector runtime your agent depends on.'

Bring user accounts into your agent without renting the runtime.

Start with a local OAuth vault and YAML connector catalog, then expose reliable CLI and MCP tools to your assistant.

Add local connectors