Brownian Code

Think Claude Code, but built specifically for crypto research.

Brownian Code is an autonomous crypto research agent that thinks, plans, and learns as it works. It performs analysis using task planning, self-reflection, and real-time market data -- decomposing complex questions into research steps, gathering live data, validating its own work, and iterating until it has a confident, data-backed answer.

Install

Quick install (macOS / Linux):

curl -fsSL https://brownian.xyz/install.sh | bash

npm:

npm install -g brownian-code

Homebrew:

brew tap brownian-xyz/brownian
brew install brownian

From source:

git clone https://github.com/brownian-xyz/brownian-code.git
cd brownian-code
bun install
bun start

Quick Start

1. Copy the example environment file and add at least one LLM API key:

cp env.example .env
# Edit .env and add your API key(s)

2. Launch the agent:

brownian

3. Ask a question:

> Compare SOL vs ETH developer activity and TVL trends over the last quarter

See QUICKSTART.md for a full first-run walkthrough.

Features

• Autonomous task planning -- Decomposes complex queries into structured research steps, executes them, and self-validates results.
• Self-reflection -- Checks its own work, detects gaps, and iterates until tasks are complete. Built-in loop detection and step limits prevent runaway execution.
• Real-time crypto data -- 227+ endpoints via Hive Intelligence covering prices, DeFi, wallets, security, NFTs, and more.
• Multi-provider LLM support -- OpenAI, Anthropic, Google, xAI, OpenRouter, Ollama (local). Switch models on the fly with /model.
• Web search + browser -- Exa/Tavily search and Playwright-based scraping for real-time web research.
• Extensible skills -- Define custom research workflows as SKILL.md files.
• Conversation history -- Persistent history with search, export, and context compaction.

Documentation

| Document | Description | |----------|-------------| | INSTALL.md | Detailed installation guide with all methods | | QUICKSTART.md | First-run setup and usage walkthrough | | docs/COMMANDS.md | All CLI commands and keyboard shortcuts | | docs/CONFIGURATION.md | Environment variables and config files | | docs/TROUBLESHOOTING.md | Common issues and fixes | | AGENTS.md | Repository guidelines for AI coding agents |

Contributing

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Open a Pull Request

Please keep pull requests small and focused. Run bun run typecheck && bun test before submitting.

License

MIT -- see LICENSE for details.