Back to directory

kimchi

A terminal coding agent CLI built on the pi-mono agent SDK that can delegate different task types to different models through a multi-model orchestrator, backed by kimchi's own LLM infrastructure.

kimchi

kimchi is a coding agent for the terminal, built on the pi-mono coding agent SDK and connected to kimchi's own LLM infrastructure rather than requiring a user-supplied API key for every provider. In its multi-model mode, an orchestrator classifies each task and delegates it to the model assigned to that role, where each role can have a single model or a pool of candidates the orchestrator picks between by tier and description; a single-model mode instead routes every task to one chosen model while keeping the same orchestration system prompt active.

The available model list is fetched at startup from kimchi's metadata service, and a user can cycle through models or open a picker to choose a specific model or multi-model mode directly from the interactive CLI.

It installs via Homebrew, a shell install script, or a PowerShell one-liner on Windows, and is written in TypeScript under the Apache-2.0 license.

More in Coding agents

See all 793
  • The City

    github.com

    A native macOS app that shows Claude Code runs as a city, with each project as a building, each team as a floor of robots, and every animation driven by a real event from the claude CLI's stream.

  • rig

    tryrig.ai

    A small, single-binary agent runtime that points at any OpenAI-compatible model, local or hosted, reads and edits a repository's code, runs its tests and remembers what it learned, with a loop implemented in a few hundred lines.

  • Jauvex

    jauvex.reindent.com

    An Electron desktop app that puts Claude Code, Codex and Grok coding-agent sessions side by side and lets you steer them by voice, with a small model handling fast routing decisions.