core vs mcp-agent
Python frameworks for building AI agents compared
core and mcp-agent are Python frameworks for building AI agents. core offers a local UI and REST API for exploring and extending conversational agents, while mcp-agent focuses on MCP server integrations and composable workflows, including optional Temporal execution.
core: Build and Extend AI Agents
Cheshire Cat AI is a Python framework and microservice for building conversational AI agents. It is aimed at learning, research, and publishing agents, with a local UI and REST API for exploring and extending them.
mcp-agent: Build Agents with Model Context Protocol
A Python framework for building LLM agents around Model Context Protocol servers and composable workflow patterns. It suits developers who need MCP integrations, multi-agent workflows, or durable execution with Temporal.
| core | mcp-agent | |
|---|---|---|
| Language | Python | Python |
| License | GPL-3.0 | Apache-2.0 |
| Stars | 3.1k | 8.6k |
| Forks | 411 | 893 |
| Last analyzed | Oct 3, 2026 | Oct 3, 2026 |
Key differences
- core is framed for learning, research, and publishing conversational agents; mcp-agent focuses on connecting LLM agents to MCP tools and composing workflows.
- core includes a browser UI and API playground; mcp-agent provides a CLI for scaffolding and deployment.
- mcp-agent offers workflow patterns such as routers and orchestrator-workers, plus asyncio or Temporal execution; core highlights plugins, function calling, vector search, and MCP in its feature description.
- core is licensed under GPL-3.0, while mcp-agent is licensed under Apache-2.0.
- core's README describes version 2 as an unstable alpha with breaking changes and advises against production use; mcp-agent describes its cloud deployment as beta and requires provider credentials and MCP configuration.
- Both are Python projects; mcp-agent examples use asynchronous patterns, and Temporal-backed durability requires a Temporal runtime and worker setup.
Choose core if you…
- want a local UI and REST API for exploring a conversational agent.
- need a framework aimed at learning, research, or publishing agents.
- want to extend an agent through plugins or explore capabilities such as function calling and vector search.
Choose mcp-agent if you…
- need agents to use tools exposed by MCP servers and manage their connections.
- want to compose workflows such as parallel map-reduce or orchestrator-workers.
- need optional Temporal execution for durable workflows, including pause, resume, and retries.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.