Everything in Condor is free and open source. We’re supported by exchanges as partners, so you can use everything without paying. Join #condor-feedback in Discord to share feedback.
Why Condor?
It’s easy to create a simple AI trading demo. But to make money over the long term, you need to:- Calculate P&L accurately across different exchanges
- Have the agent actually learn from what it’s doing
- Run many agents that operate independently but also work together
Key Features
Autonomous Trading
Deploy AI agents that observe, reason, and execute independently—no manual intervention
Multi-Agent Support
Run dozens of agents on shared accounts with isolated P&L tracking per agent
50+ Exchanges
Trade on CEX, DEX, and blockchain networks with standardized execution
LLM Integration
Connect to Claude, Gemini, GPT, or local models via Agent Client Protocol
Full Observability
Every tick captured: prompt, reasoning, tool calls, and results
Deterministic Routines
Move expensive computations out of the LLM into reusable Python code
How It Works
Condor uses a two-server architecture separating the agentic layer (LLM reasoning) from the execution layer (Hummingbot API):| Component | Role |
|---|---|
| Condor Server | LLM integration, Telegram interface, agent state management |
| Hummingbot API | Data collection, trade execution, bot orchestration |
Getting Started
Motivation
Understand why Condor is architected this way
Quickstart
Deploy Condor and execute your first trade
API Reference
Explore all Hummingbot API endpoints
GitHub
Source code and issues