Your AI coding agent, in the terminal. Fast. Powerful. Written in Rust.
A terminal-based AI coding agent that helps you read, write, edit code, execute commands, and reason through complex problems — all through natural language.
- 🔌 Multi-LLM — OpenAI, Anthropic, Google AI, Azure OpenAI, custom endpoints
- 🛠️ Built-in Tools — Bash, file edit, glob search, JSON edit, code knowledge graph (tree-sitter)
- 🤖 Multi-Agent — Coordinator, Coder, Reader, Reviewer for complex tasks
- 🧠 Thinking Models — Native Claude extended thinking support
- 💾 Smart Memory — Intelligent conversation compression within token limits
- ⚡ Skill System —
/slashcommands: git commit, deep research, web scraping, PDF analysis - 🖥️ Interactive TUI — Syntax highlighting, file search, rich terminal UI
git clone https://github.com/21pounder/terminal-coding-agent.git
cd terminal-coding-agent
cargo install --path cliRequires: Rust 1.75+ and an API key from any supported LLM provider
Create tca.json in your project root or ~/.config/tca/config.json:
{
"protocol": "anthropic",
"api_key": "your-api-key",
"base_url": "https://api.anthropic.com",
"model": "claude-sonnet-4-20250514"
}Or use environment variables:
export ANTHROPIC_API_KEY=your-keySupported: openai · anthropic · google_ai · azure_openai
tcarust> Fix the token refresh bug in src/auth.rs
> Add input validation to the signup form
> Explain how the routing system works
core/ # Core library — agent loop, LLM clients, tool system
├── agent/ # Agent core, prompt, config, conversation management
├── agents/ # Multi-agent (coordinator, coder, reader, reviewer)
├── llm/ # LLM clients (OpenAI, Anthropic)
└── tools/ # Built-in tool definitions
cli/ # CLI — TUI, interactive mode, tool implementations
├── interactive/# Terminal UI components
├── tools/ # bash, edit, glob, json_edit
└── config/ # Configuration loader
MIT OR Apache-2.0
