Back to directory

codehamr

A minimal, standalone terminal coding agent that works through four tools, bash, read_file, write_file, and edit_file, connecting to any local model server or API provider with no project account or subscription required.

codehamr

codehamr is a small coding agent, released under the MIT license, that reads a project, makes changes, runs checks, and reports back through just four tools and two slash commands, with one embedded system prompt and one configuration file. It connects to a local model server such as Ollama or to a hosted API such as OpenRouter, configured through named profiles that specify a model ID, base URL, optional key, and context window; a key can be given literally or read from an environment variable at connection time.

Shell commands execute with the user's own permissions and require a POSIX shell, so on Windows it needs to run inside WSL2 or a Linux devcontainer; the tool itself provides no sandboxing, so isolation is left to a devcontainer or virtual machine when needed. Token counts used for packing context are estimates, and the client reserves space for the system prompt, tool schemas, and response before filling the rest with recent conversation history.

It installs with a shell script on Linux and macOS or a batch script on Windows, downloading prebuilt binaries from GitHub releases.

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.