glincker/awesome-selfhosted-ai

Curated list of self-hosted open-source AI: local LLMs, chat UIs, agents, MCP servers, RAG, speech and image generation. Health-checked weekly, dead projects removed automatically

Summary

glincker/awesome-selfhosted-ai is a curated directory of open-source AI software that can run on your own hardware or server. It organizes 303 projects into categories such as local LLM runtimes, chat interfaces, coding agents, RAG, MCP servers, speech, and image generation, with a weekly GitHub API check that moves archived, deleted, or long-inactive projects to UNMAINTAINED.md.

What it is useful for

The repository helps engineers discover alternatives before committing to a particular self-hosted AI stack. Instead of searching for each component separately, engineers can compare projects by workload:

• Serving models: Ollama provides a local model runner with a CLI and HTTP API, while LocalAI exposes text, image, and audio models through an OpenAI-compatible server.
• Building chat applications: AnythingLLM supports document chat with local or cloud models and built-in RAG; LibreChat provides a Docker-based multi-provider interface with plugins, file uploads, and agents.
• Adding coding workflows: The list includes Cline, OpenHands, SWE-agent, and terminal-based tools such as Codex CLI and opencode.
• Connecting agents to tools: Its MCP section covers gateways, memory, browser inspection, and tools for terminals, files, and code search.

It is also useful when planning a homelab or private server deployment because projects are grouped by functional area rather than product type. The catalog includes a separate “Source-Available” section, so engineers should still check each project’s license rather than assuming every listed tool has the same distribution terms.

How engineers can use it

Start with the README’s category index and narrow the search to the capability needed. For example, for local document chat, compare AnythingLLM’s built-in RAG with LibreChat’s Docker deployment. For a model server, compare Ollama with the OpenAI-compatible APIs offered by LocalAI, llama.cpp-based tools, or larger serving platforms.

Then use LIVE.md to compare stars, last push activity, and license for each candidate. Review the linked project’s own documentation for supported hardware, dependencies, and deployment details before choosing it.

The repository README does not provide setup commands for the listed tools, and the catalog itself is not an installer or compatibility guarantee. For installation instructions, consult the README of the specific project you select. Also check the maintenance records, since the catalog explicitly treats long-term inactivity, deletion, or archival as reasons for moving an entry out of the main list.