The AI coding agent built for the terminal.
Available as a terminal interface. Supports 75+ LLM providers, including local models.
- [*] Native TUI—A responsive, native, themeable terminal UI.
- [*] LSP enabled—Automatically loads the right LSPs for the LLM.
- [*] Multi-session—Run multiple agents in parallel on the same project, coordinated by a Supervisor.
- [*] GitHub Copilot—Log in with GitHub to use your Copilot account.
- [*] ChatGPT Plus/Pro—Log in with OpenAI to use your ChatGPT Plus or Pro account.
Built into hordecode
A persistent session that routes work across a fleet of background agents as tracked child sessions and reports their results back to you — you're never left babysitting subagents.
Every task runs through the matching skill before implementation starts — 16 skills covering plan, TDD, debug, review and verify, including a verification skill that asks for evidence before work is called done.
A dedicated pipeline for investigation and deep research, separate from the build pipeline — 16 of its own skills, run by a dedicated research agent.
Every active agent, subagent, and cron job shows up in the sidebar in real time, with status and elapsed time — full visibility into what's actually running, not a black box.
Permission rules and a sandboxed workflow engine, built into the runtime rather than added as plugins.
7 built-in memory tools (add, recall, update, session, status, dump, gc). Recall is keyword search (BM25) by default. Semantic search is opt-in with memory.embedding: a one-time download of about 133 MB, available on linux-x64. No plugin and no separate API key.
