A collection of lightweight CLI tools for working with local and cloud LLMs. Each tool is small, focused, and designed to compose cleanly in scripts and pipelines.
| Project | Description |
|---|---|
| lite-llm | Lightweight CLI for OpenAI-compatible LLM APIs. Supports batch mode, structured output, streaming, and prompt-injection protection. |
| lite-rag | CLI-based RAG tool for Markdown documents. Indexes files into a local DuckDB vector store and answers natural-language questions via a local LLM. |
| lite-switch | Natural language classifier for shell pipelines. Reads free-form text from stdin and outputs the best-matching tag via an OpenAI-compatible LLM. |
| lite-eml | EML parser for shell pipelines. Extracts headers and body from .eml files as structured JSONL, with full charset decoding (ISO-2022-JP, Shift_JIS, etc.). |
| lite-msg | Outlook MSG parser for shell pipelines. Extracts headers and body from .msg files as structured JSONL (same schema as lite-eml). |
- Small and focused — each tool does one thing well and composes with others via stdin/stdout.
- Local-first — works with local LLMs (LM Studio, Ollama) out of the box; cloud APIs are opt-in.
- Scriptable — no interactive UI required; all configuration is file- or environment-based.
All projects in this series follow shared conventions documented in CONVENTIONS.md: config file format, CLI flag naming, Makefile targets, release process, package structure, and more.