Session: Codename Goose: An Open Source Local AI Agent Built on MCP
AI coding assistants are evolving fast, but most are limited to generating code. Codename Goose, an open source AI agent, takes it further by actually running commands, fixing code, and interacting with your development environment like a true assistant.
Unlike cloud-based AI tools that rely on proprietary APIs, Goose runs locally on your machine and allows you to use the LLM of your choice, giving you full control over its actions and integrations. It’s built on the Model Context Protocol (MCP), an open standard that connects Goose to 1,700+ extensions giving it the ability to fetch data, interact with your databases, IDEs, GitHub repos, Google apps, and much much more.
In this session, we’ll explore:
- How Goose turns LLMs into actionable agents, capable of more than just code suggestions.
- How Goose is infinitely extensible, letting you integrate it with any LLM and any API.
- Real-world use cases.
Goose is local-first, developer-first, and open source. Come see it in action and learn how to build, extend, and customize your own workflows.
