noni-chan (Agentic Coding CLI)

To install dependencies:

bun install

To run:

bun dev

This CLI renders a terminal UI using Ink (React for CLIs).

Model providers

The agent will pick the first configured provider in this order (env vars override config):

• Claude (Anthropic): set ANTHROPIC_API_KEY
• Gemini (Google): set GEMINI_API_KEY (or GOOGLE_API_KEY)
• OpenAI: set OPENAI_API_KEY (optional OPENAI_BASE_URL, OPENAI_MODEL)
• Ollama (hosted/self-hosted): set OLLAMA_BASE_URL and OLLAMA_MODEL
  • Optional: OLLAMA_API_KEY (only if your hosted Ollama requires auth)

Store keys in a config file (recommended for npx)

noni-chan stores settings in a cross-platform location:

• Windows: %APPDATA%\\noni-chan\\settings.json
• macOS: ~/Library/Application Support/noni-chan/settings.json
• Linux: $XDG_CONFIG_HOME/noni-chan/settings.json (or ~/.config/noni-chan/settings.json)

Set values:

noni-chan config set anthropicApiKey "YOUR_ANTHROPIC_KEY"
noni-chan config set geminiApiKey "YOUR_GEMINI_KEY"
noni-chan config set openaiApiKey "YOUR_OPENAI_KEY"
noni-chan config set openaiModel "gpt-4o-mini"
noni-chan config set ollamaBaseUrl "https://your-ollama-host.example.com"
noni-chan config set ollamaModel "llama3.1:8b"

Show current config:

noni-chan config show

Hosted Ollama example

PowerShell:

$env:OLLAMA_BASE_URL="https://your-ollama-host.example.com"
$env:OLLAMA_MODEL="llama3.1:8b"
# optional if your host requires it:
$env:OLLAMA_API_KEY="YOUR_TOKEN"

bun dev --headless "Hello! Say hi in one sentence."