# llms.py

llms.py is a self-hosted ai app. A lightweight CLI, API and ChatGPT-style web interface for chatting with multiple large language model (LLM) providers, running entirely offline with all data kept private in browser storage. It is open source and self-hosted, installable on TrueNAS.

Canonical page: https://ourlocalhost.com/app/4390/llms-py

## Description

A lightweight CLI, API and ChatGPT-style web interface for chatting with multiple large language model (LLM) providers, running entirely offline with all data kept private in browser storage.

## Platform availability

Available on 4 of 10 platforms: TrueNAS, Mac, Windows, Linux.

## Desktop and other devices

Mac, Windows, Linux.

## Health score

66/100 (Good), based on 4 of 5 factors.
- Maintained: 100 (last commit 27 days ago)
- Popular: 56 (628 GitHub stars)
- Easy to install: 55 (4 install platforms, install notes)
- Security check: 20 (1 serious finding, 5 minor, CI, tests)

## Security check

1 serious finding: Uses zip 4.6.1, which has a known critical vulnerability.
Automated checks on commit cc90dc2, 30 Oct 2026. Not an audit.

## Links

- GitHub: https://github.com/ServiceStack/llms
- Documentation: https://llmspy.org/docs
- desktop: https://github.com/ServiceStack/llms/releases/latest
- TrueNAS: https://apps.truenas.com/catalog/llmspy/
- Official website: https://llmspy.org/

## Common questions

### How do I start the web UI?

Install with pip install llms-py, set provider API keys, run llms --serve 8000, and open the UI on local port 8000.

### Where is chat data stored?

The docs say it is privacy-first and that all data is stored locally in the browser.

### Can it automatically switch between providers?

Yes, it includes intelligent routing with automatic failover between providers.

### Is there a Docker deployment option?

The docs describe llms.py as Docker-ready with pre-built images, and Docker Deployment is listed in the documentation.

### Are there sign-in options?

The documentation lists Authentication, Credentials, ChatGPT Sign-In, and GitHub Auth configuration.

### Does it support image and audio workloads?

The site says it supports text, image, and audio, with image generation from several providers plus text-to-speech and voice-to-text input.

---
Source: Our Local Host (https://ourlocalhost.com/app/4390/llms-py), the self-hosted and open-source app directory.
