rustchain

Keep your existing workflows. RustChain runs them all.

RustChain consumes workflows from the tools you already use:

You don't have to rewrite anything. Point RustChain at your existing files and it generates executable missions.

# Your existing LangChain script
rustchain transpile lang-chain my_agent.py -o mission.yaml

# Now run it with RustChain
rustchain run mission.yaml

Why migrate?

How Transpilation Works

RustChain's transpiler parses your existing workflow definitions and generates equivalent RustChain missions:

Your Workflow          RustChain              Output
─────────────          ─────────              ──────
LangChain.py      ──▶  Transpiler        ──▶  mission.yaml
Airflow DAG            (parses & converts)    (executable)
GitHub Actions
K8s manifest
Docker Compose

What gets converted

• Steps/Tasks → Mission steps with proper types
• Dependencies → depends_on relationships preserved
• Configuration → Parameters mapped to RustChain equivalents
• Secrets/Env vars → Environment variable references preserved

Auto-detection

Don't know the format? RustChain figures it out:

rustchain transpile auto my_workflow.py
# Detects: LangChain Python
# Output: my_workflow.yaml

Supported conversions

Quick Start

1. Install

git clone https://github.com/Michael-A-Kuykendall/rustchain
cd rustchain
cargo build --release --features "cli,tools,llm"

2. Convert an existing workflow

# Have a LangChain script?
rustchain transpile lang-chain your_script.py -o mission.yaml

# Have an Airflow DAG?
rustchain transpile airflow your_dag.py -o mission.yaml

# Not sure what format?
rustchain transpile auto your_file.py -o mission.yaml

3. Validate and run

# Check it first
rustchain mission validate mission.yaml

# Dry run (no side effects)
rustchain run mission.yaml --dry-run

# Execute
rustchain run mission.yaml

Native Mission Format

If you want to write missions directly (instead of converting), use YAML:

name: "Data Pipeline"
version: "1.0"
steps:
  - id: "fetch_data"
    name: "Fetch from API"
    step_type: "http_request"
    parameters:
      url: "https://api.example.com/data"
      method: "GET"

  - id: "process"
    name: "Process with LLM"
    step_type: "llm"
    depends_on: ["fetch_data"]
    parameters:
      provider: "ollama"
      model: "llama2"
      prompt: "Summarize: ${fetch_data.output}"

  - id: "save"
    name: "Save results"
    step_type: "create_file"
    depends_on: ["process"]
    parameters:
      path: "output.txt"
      content: "${process.output}"

CLI Reference

rustchain run <file>           Execute a mission
rustchain transpile <cmd>      Convert from other formats
rustchain mission <cmd>        Mission management
rustchain safety <cmd>         Safety validation
rustchain tools <cmd>          Tool management
rustchain audit <cmd>          Audit queries
rustchain policy <cmd>         Policy management
rustchain config <cmd>         Configuration
rustchain interactive          Interactive mode

See rustchain --help for full details.

Documentation

• Installation
• Quick Start
• Transpilation Guide
• CLI Reference
• Examples

Requirements

• Rust 1.70+: Required for compilation
• Optional: Ollama or compatible LLM backend for AI features

License

Dual-licensed under MIT or Apache-2.0, at your option.

See LICENSE-MIT and LICENSE-APACHE.

Contributing

RustChain is open source but not open contribution. See CONTRIBUTING.md.

Security

To report vulnerabilities, see SECURITY.md.