LANCompute v0.1.0 — LM Studio/Ollama integration, CLI helper, tests, and docs

LANCompute v0.1.0 — LM Studio/Ollama integration, CLI helper, tests, and docs

Summary

• Adds a minimal OpenAI-compatible workflow to exercise a local LLM service (LM Studio API server or Ollama).
• Introduces a tiny CLI for quick prompts and model discovery.
• Includes an integration test to verify /v1/models and /v1/chat/completions.
• Adds environment templates, updates docs, and improves developer ergonomics.

Highlights

• CLI helper: scripts/lmstudio_chat.py
  • Lists models and sends quick chat prompts to an OpenAI-compatible endpoint.
  • Reads .env automatically and respects LM_STUDIO_BASE_URL.
  • Works out-of-the-box against http://127.0.0.1:1234 (override via env or flag).
• Makefile targets
  • make models lists models via the CLI helper.
  • make chat MODEL=<id> PROMPT="<text>" sends a quick prompt.
  • make setup upgrades pip and ensures requests is present.
• Integration test: tests/test_lmstudio_integration.py
  • Discovers available models, prefers lightweight IDs, and runs a short chat completion.
  • Environment overrides: LM_STUDIO_BASE_URL, LM_STUDIO_TEST_MODEL, LM_STUDIO_TEST_PROMPT.
• Configuration and docs
  • .env.example for public-safe defaults.
  • .envrc auto-activates .venv and loads .env (with direnv).
  • .gitignore ignores .env and .direnv/.
  • README adds integration guide, CLI usage, env setup, and test instructions.

Getting Started

python -m venv .venv
source .venv/bin/activate
pip install -U pip requests pytest
# optional: direnv allow
cp .env.example .env
# set LM_STUDIO_BASE_URL (default: http://127.0.0.1:1234)

Usage

python scripts/lmstudio_chat.py --list-models
make models
python scripts/lmstudio_chat.py --model mistral:latest --prompt "Give me one fun fact."
make chat MODEL=mistral:latest PROMPT="Give me one fun fact."

Integration Test

pytest -q tests/test_lmstudio_integration.py

Environment overrides:

• LM_STUDIO_BASE_URL — API base URL
• LM_STUDIO_TEST_MODEL — preferred model id (optional)
• LM_STUDIO_TEST_PROMPT — short test prompt (optional)

Security Notes

• Only expose your LLM service to a trusted network.
• Prefer binding locally (127.0.0.1) and use SSH tunnels for development.
• If you bind 0.0.0.0, ensure you control access (firewall/VLAN/reverse proxy).

Changelog

• Add scripts/lmstudio_chat.py CLI for models and chat
• Add Makefile targets: models, chat, setup
• Add tests/test_lmstudio_integration.py with env overrides
• Add .env.example; load .env in CLI and tests
• Update .envrc to source .env
• Update .gitignore to exclude .env and .direnv/
• Expand README with integration guide and usage