mini-coder

Minimalistic terminal coding agent, with an emphasis on giving you control of what is sent to the model, and visibility into the model's actions.

Beautiful TUI with Tokyo Night colors, more to come soon.

We try to render diffs when the agent uses bash to edit, only for the current working directory; changes outside it aren't tracked.

Install

npm i -g mini-coder

Configuration

You can use /login and /provider to configure in the TUI. Changes are saved to your config.

Global config lives at ~/.config/mini-coder/config.json ($XDG_CONFIG_HOME/mini-coder/config.json if set). You can also configure your provider and model there:

{
  "provider": "opencode-go",
  "model": "deepseek-v4.1-flash"
}

Defaults for the rest:

{
  "sessionsDir": "./sessions",
  "systemPrompt": "",
  "discoverAgentFiles": true,
  "skillsDirs": [],
  "tools": ["edit", "read", "bash"],
  "thinkingEffort": "medium",
  "customProviders": []
}

• provider / model — any model from the pi-ai catalog. Requires a discoverable api key from the environment (like OPENCODE_API_KEY). When neither is set, the agent launches; sending a message reports that no model is configured.
• sessionsDir — where append-only session JSONL files are written. (no relative paths for now, absolute paths only).
• systemPrompt — the base of the system prompt. Skills and agent files, if enabled, are appended after it.
• discoverAgentFiles — when true, appends AGENTS.md / CLAUDE.md found in ~/.agents and the working directory to the system prompt.
• skillsDirs — directories of skills to advertise (<dir>/<skill>/SKILL.md with name: and description: frontmatter). Paths, not contents, go to the model.
• tools — which of edit, read, bash the agent gets. Tool arguments are untrusted and validated at this boundary regardless.
• thinkingEffort — minimal, low, medium, high, xhigh, or max for models that support reasoning.

Local and custom models

Hosted and local models are equals. Add unlisted or self-hosted models through customProviders — the four wire APIs come from pi-ai, and auth resolves from your environment via envKeys (no local model needs it):

{
  "customProviders": [
    {
      "id": "ollama",
      "name": "Ollama",
      "baseUrl": "http://localhost:11434/v1",
      "api": "openai-completions",
      "models": ["qwen3-coder:30b", "gpt-oss:20b"]
    }
  ]
}

api is openai-completions, openai-responses, anthropic-messages, or google-generative-ai. Then point provider and model at it.