Back to directory

Magma

A TypeScript framework for building and deploying conversational AI agents quickly, with a tool and middleware system and one-command hosted deployment.

Magma

Magma is a framework for creating AI agents without long chains of abstractions: a developer defines the logic they want an agent to have directly, decorates methods as tools the agent can call, and runs the result through a small API. Tools give an agent the ability to perform actions and must each return a string, and middleware methods run custom logic before or after a completion or a tool call, such as logging, validation or sanitizing data, and can throw errors that redirect the conversation flow in a defined way.

Agents can use any of several AI providers, including OpenAI, Anthropic and Groq, and can be deployed in seconds on a separate hosted deployment platform or self-hosted. The project ships with a demo agent used for user research, built with the framework.

It installs as an npm package.

More in Agent frameworks

See all 902
  • Recurse

    recurse.run

    A serverless harness that lets a coding agent design, test, and deploy its own specialist sub-agents as tools, MCP servers, or bots, defined by a schema-validated manifest.

  • ProofForge

    shulgin.is-a.dev

    An AI-agent pipeline that decomposes a mathematics problem, proves each piece, and formalizes the result in Lean 4 and Mathlib until the Lean kernel accepts it, with six of its proofs already merged into Google DeepMind's formal-conjectures repository.

  • local-coder

    github.com

    Builds a local coding-agent team on top of Ollama and OpenCode by detecting the machine's hardware, recommending models for seven specialist roles, and testing whether each assigned model can actually use tools, edit a repository and pass verification.