Human-Readable First
Always exports something you can skim in 2 minutes.
Built for Real Workflows
Use it when switching tools, agents, or worktrees.
Architecture Handoffs
Export the current state for a deeper reasoning pass with another model.
Research Handoffs
Keep sources, decisions, and constraints bundled for seamless context transfer.
Multi-Agent Execution
Assign tasks across models and track who is doing what.
Local Model Fallback
Use Ollama or LMStudio when you need privacy or offline access.
The Workflow
Simple three-step process: import, extract, export.
Import
Bring in a conversation from any AI tool or IDE.
Extract
Pull out decisions, code blocks, and next steps automatically.
Export
Generate a clean markdown handoff for the next tool or model.
Get Started in Seconds
Install via pip or pipx and start orchestrating.
pip install orchestrator-cli
Install globally with pip for system-wide access.
pipx install orchestrator-cli
Recommended: isolated environment with pipx.
orchestrator init
Initialize a new project with default templates.
Built for Developers Who Ship
No SaaS lock-in. No mandatory cloud. Just a CLI that helps you move faster.
Works Locally
CLI-first design with no mandatory cloud dependency.
Open Source
MIT licensed. Fork, customize, and contribute.
Model Agnostic
Works with any LLM provider or local model.