Skip to Content

Home MCP Servers API & SDKs

CTRL K

CTRL K

Home
Get Started
About Arcade
- Get an API key
- Connect Arcade docs to your IDE
- Call a tool in your agent
- Call a tool in your IDEs, MCP clients, or agents
Example agents
Toolkits
Glossary
FAQ
Changelog
Guides
- Set up secure multi-user auth for your app
- Placeholder for Quickstarts of various observability platforms
Learn
- Why and how do agents call tools
- Intro to TAG
How do auth and secrets work
- - Background agents + secrets
  - Agent apps + auth
- - Supervisor architecture
API Reference
API
Arcade MCP (MCP server SDK)
Blog
Status

Question? Give us feedback Edit this page on GitHub →

Home Create toolsEnsure tools work with the newest modelsRun existing eval and observe outcome with new model

Run existing eval and observe outcome with new model

Coming soon!

Last updated on December 11, 2025

Write custom toolkit Modify tool to work well with new model

Making AI take action

Project

GitHub
GitHub Container Registry

Resources

Documentation
Releases
Examples

Feedback

Discord
Report an issue on GitHub

© 2024 — present Arcade AI, Inc.

Run existing eval and observe outcome with new model | Arcade Docs