Self-hosted alternatives to Microsoft Copilot
Want AI chat assistants (run models on your own hardware) without handing your data to Microsoft Copilot? These 14 self-hosted apps run on your own server (most are open source; licenses are listed below). Cheapest estimated hosting starts at about $40/month.
About Microsoft Copilot: Microsoft's AI assistant.
| App | Health | Stars | License | Docker | RAM (est.) | From |
|---|---|---|---|---|---|---|
| Ollama | Active | 186k | MIT | ✓ | 8 GB | $40/mo |
| Open-WebUI | Active | 157.8k | BSD-3-Clause | ✓ | 8 GB | $40/mo |
| LobeHub | Active | 84k | ⊘ Proprietary | ✓ | 8 GB | $40/mo |
| AnythingLLM | Active | 68k | MIT | ✓ | 8 GB | $40/mo |
| LocalAI | Active | 50.2k | MIT | ✓ | 8 GB | $40/mo |
| LibreChat | Active | 46.8k | MIT | ✓ | 8 GB | $40/mo |
| Vane | Active | 38k | MIT | ✓ | 8 GB | $40/mo |
| Onyx Community Edition | Active | 33.2k | MIT | ✓ | 8 GB | $40/mo |
| Local Deep Research | Active | 9.5k | MIT | ✓ | 8 GB | $40/mo |
| Agenta | Active | 4.9k | MIT | ✓ | 8 GB | $40/mo |
| Khoj | Maintained | 38k | AGPL-3.0 | ✓ | 8 GB | $40/mo |
| LLM Harbor | Active | 3.3k | Apache-2.0 | ✓ | 8 GB | $40/mo |
| GoModel | Active | 1.3k | MIT | ✓ | 8 GB | $40/mo |
| LLMKube | Active | 252 | Apache-2.0 | ✓ | 8 GB | $40/mo |
The options in detail
1. Ollama
Get up and running with Llama 3.3, DeepSeek-R1, Phi-4, Gemma 3, and other large language models.
Active Last commit 2026-11-05. Latest release v0.40.15-rc0 (2026-11-03). Needs roughly 8 GB RAM. How to run it →
2. Open-WebUI
User-friendly AI Interface, supports Ollama, OpenAI API.
Active Last commit 2026-11-03. Latest release v0.13.0 (2026-11-03). Needs roughly 8 GB RAM. How to run it →
3. LobeHub
Modern design AI chat framework supporting multiple AI providers, one click install MCP Marketplace and Artifacts / Thinking.
Active Last commit 2026-11-05. Latest release v2.3.2-canary.2 (2026-11-05). Needs roughly 8 GB RAM. How to run it →
4. AnythingLLM
All-in-one desktop & Docker AI application with built-in RAG, AI agents, No-code agent builder, MCP compatibility, and more.
Active Last commit 2026-11-05. Latest release v1.19.0 (2026-10-26). Needs roughly 8 GB RAM. How to run it →
5. LocalAI
Run your AI models locally and generate images and audio (alternative to OpenAI and Claude).
Active Last commit 2026-11-05. Latest release v4.12.1 (2026-10-23). Needs roughly 8 GB RAM. How to run it →
6. LibreChat
Enhanced ChatGPT-compatible AI chat interface supporting multiple AI providers, with multi-user auth, message search, and plugin support.
Active Last commit 2026-10-31. Latest release v0.8.9-rc1 (2026-10-26). Needs roughly 8 GB RAM. How to run it →
7. Vane
AI-powered search engine (alternative to Perplexity AI).
Active Last commit 2026-09-01. Latest release v1.12.2 (2026-04-10). Needs roughly 8 GB RAM. How to run it →
8. Onyx Community Edition
Chat UI that works with any LLM. It comes loaded with advanced features like agents, web search, RAG, MCP, deep research, Connectors to 40+ knowledge sources, and more.
Active Last commit 2026-11-05. Latest release helm/onyx-0.9.31 (2026-11-03). Needs roughly 8 GB RAM. How to run it →
9. Local Deep Research
AI-powered deep research tool with multi-source search (arXiv, PubMed, web), PDF text extraction, and encrypted local storage.
Active Last commit 2026-11-04. Latest release v1.10.7 (2026-08-28). Needs roughly 8 GB RAM. How to run it →
10. Agenta
LLMOps platform for prompt management, LLM evaluation, and observability. Build, evaluate, and monitor production-grade LLM applications with collaborative prompt engineering.
Active Last commit 2026-11-04. 10+ releases in the past year. Latest release v0.128.6 (2026-11-04). Needs roughly 8 GB RAM. How to run it →
11. Khoj
Your AI second brain. Get answers from the web or your docs. Build custom agents, schedule automations, do deep research. Turn any online or local LLM into your personal, autonomous AI.
Maintained Last commit 2026-08-02. Latest release 2.0.0-beta.28 (2026-03-26). Needs roughly 8 GB RAM. How to run it →
12. LLM Harbor
Containerized LLM toolkit. Run LLM backends, APIs, frontends, and additional services via a concise CLI.
Active Last commit 2026-10-31. 10+ releases in the past year. Latest release v0.5.19 (2026-10-31). Needs roughly 8 GB RAM. How to run it →
13. GoModel
AI gateway written in Go with a unified OpenAI-compatible API for multiple LLM providers, USD cost tracking, budgets, usage analytics, guardrails, caching, and an admin dashboard.
Active Last commit 2026-11-04. 10+ releases in the past year. Latest release v0.2.1 (2026-10-30). Needs roughly 8 GB RAM. How to run it →
14. LLMKube
Kubernetes operator for LLM inference with pluggable runtimes (llama.cpp, vLLM, TGI, Ollama, vllm-swift), multi-GPU sharding, NVIDIA CUDA + Apple Silicon Metal support, and OpenAI-compatible API.
Active Last commit 2026-11-04. 10+ releases in the past year. Latest release v0.10.16 (2026-11-02). Needs roughly 8 GB RAM. How to run it →
Why self-host instead of Microsoft Copilot?
- Own your data — it lives on a server you control, with backups you decide on.
- Predictable cost — a small VPS often replaces per-seat pricing for a whole team or family.
- No lock-in — open-source licenses let you keep running the software even if a vendor disappears.
The trade-off: you are responsible for updates, backups and security. Pick an app marked Active (a commit in the last 90 days) and with Docker support for the easiest ride.
Browse related categories
Other alternatives guides
Microsoft Copilot is a trademark of its owner; this site is not affiliated with it. Listings come from the community-maintained awesome-selfhosted list.