Fast-agent is a Python-based framework for building and evaluating agents and MCP workflows, with extensive CLI and MCP support. It emphasizes an interactive shell, MCP servers, and agent tooling across multiple providers.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
fast-agent is a Python-based framework to create, manage, and run autonomous agents and their workflows, with MCP server support and a CLI-first interface. It supports interactive shell mode, TUI-like features, and integration with various model providers through MCP and ACP.
How it works
The repository provides tooling to define Agents, Chains, Makers, and Workflows (Parallel, Human Input, etc.), and to connect to MCP servers via OAuth-enabled transports. It includes an interactive session mode and commands to scaffold agents, run agents, and manage model configurations. It also documents a declarative syntax for agent definitions and workflow composition, and demonstrates integration patterns with MCP servers defined in fast-agent.yaml.
Getting started
To begin, install the uv package manager and Fast Agent MCP:
uv pip install fast-agent-mcp # install fast-agent!
fast-agent go # start an interactive session
fast-agent go --url https://hf.co/mcp # with a remote MCP
fast-agent go --model=generic.qwen2.5 # use ollama qwen 2.5
fast-agent go --pack analyst --model haiku # install/reuse a card pack and launch it
fast-agent scaffold # create an example agent and config files
uv run agent.py # run with default_model or FAST_AGENT_MODEL configured
uv run agent.py --model='gpt-5.4-mini?reasoning=low' # specify a model
uv run agent.py --transport http --port 8001 # expose as MCP server (server mode implied)
fast-agent quickstart workflow # create "building effective agents" examples
With no configured model, an interactive run opens the model picker. Automation and server runs must pass --model, set FAST_AGENT_MODEL, configure default_model, or declare a model on the agent.
For packaged starter agents, use fast-agent go --pack <name> --model <model>.
# Install fast-agent
uv tool install -U fast-agent-mcp
# Run fast-agent with opus, shell support and built-in subagents
fast-agent --model opus -xx
Model selection examples and MCP schemes are described throughout the docs, including how to connect to MCP servers and how to run agents with various model strings.
Recent releases
Latest releases include:
- v0.9.30 (2026-08-02): Deepseek, ZAI and Moonshot providers added; note Deepseek migrated to Responses and only models via that endpoint are supported (-flash-0731).
- v0.9.20 (2026-07-21): Signals; Anthropic/XAI optimizations; environment/patch fixes.
- v0.9.15 (2026-07-18): Background processing improvements, workspace handling updates, and tests.
- v0.9.7 (2026-07-13): ATIF v1.7 support, GPT-5.6 enhancements, native export.
- v0.9.5 (2026-07-10): GPT-5.6, Muse Spark, Grok 4.5 integrations; new presets for sol/terra/luna.
Full changelog is linked in the repository.
Traction
Stars: 3880. Forks: 424. Open issues: 24. Language: Python. License: Apache-2.0.
Behind the repo
Not provided in the README excerpt.
Caveats
License: Apache-2.0. Creation date: 2025-01-18. Last push: 2026-08-04. Repository topics include a2a, acp, agent, agent-framework, agent-skills, cli, mcp, mcp-client, mcp-server, python, skills, tui.