Quick Start

Install

Choose the install path that matches your environment. Use these options when you want platform-specific setup or Docker install/run.

Requires Python 3.11+.

Install uv first using the official docs: Astral uv installation guide.

uv tool install "fastfold-agent-cli[all]" --python 3.11

uv tool install "fastfold-agent-cli[win_build]" --python 3.11

# Install image
docker pull fastfold/fastfold-agent-cli:latest

# Run
docker run --rm -it fastfold/fastfold-agent-cli:latest
# Pin a specific release:
docker run --rm -it fastfold/fastfold-agent-cli:0.0.57

Fastfold Agent CLI (fastfold) is an agentic research environment for drug discovery and computational biology. Think of it as a coding agent, but for biology. You ask questions in natural language, and it plans and executes multi-step workflows with built-in tools, skills, and cloud integrations. Our mission is to bring the best tools to scientists wherever they work: on the cloud, on local compute, university HPC, or inside the enterprise. Under the hood it runs on a Deep Agents (LangChain / LangGraph) agentic loop with Programmatic Tool Calling (PTC) and progressive skill discovery. Many of its tools and prompts trace back to CellType, which reports a state-of-the-art 90% on BixBench-Verified-50; see Acknowledgements for credits.

Why Fastfold Agent CLI

• 190+ domain tools across target discovery, chemistry, expression, viability, safety, and structure workflows.

• Installable skills to discover and run reusable workflows (fastfold skills find, fastfold skills add ...).

• Any model setup including Anthropic, OpenAI, and local/open models through OpenAI-compatible backends (like Ollama, Unsloth, oMLX, DS4, llama.cpp, and LM Studio).

• Compute flexibility for heavy workflows like folding, protein design, and MD across Fastfold Cloud or your own infrastructure.

• Efficient agent runtime built on Deep Agents (LangGraph) with progressive skill discovery and Programmatic Tool Calling, so large skill/tool catalogs stay fast and token-light.

• PyPI: fastfold-agent-cli

• GitHub: fastfold-ai/fastfold-agent-cli

Upgrading the Fastfold Agent CLI

Keep your CLI updated to get new tools and fixes.

/upgrade

uv tool install "fastfold-agent-cli[all]" --python 3.11 --upgrade

uv tool install "fastfold-agent-cli[win_build]" --python 3.11 --upgrade

Check your installed version:

fastfold --version

Command guides

This CLI section is split into child pages so each command area is easier to scan:

• Setup and local models (fastfold setup, provider config, and compatible profile templates)
• Usage and quick start (fastfold, single prompts, fastfold doctor, fastfold agent fork, resume)
• Interactive commands (/help, /model, /model-manager, /tasks, /upgrade, and more)
• Skills commands (fastfold skills ... and /skills...)
• Data and reports (fastfold data ..., fastfold report ..., /export-share)
• Tool categories (domain map of available tools)
• Sessions and persistence (resume/history files and commands)
• Troubleshooting (common errors and fixes)

Recommended first steps

If you are setting up for the first time, follow this order:

1. Install from this page.
2. Configure keys and provider in Setup and local models.
3. Run your first query in Usage and quick start.

Benchmarks

Coming soon. We're crafting a comprehensive benchmark focused on industry-specific drug-discovery and computational-biology use cases: the multi-step, tool-heavy workflows scientists actually run, measured across multiple model backends. We're looking for contributors: if you have a real-world use case you'd like represented, or want to add tasks, datasets, or scoring rubrics, open an issue or reach out on Slack.