AnythingLLM
All-in-one desktop & Docker AI application with built-in RAG, AI agents, No-code agent builder, MCP compatibility, and more.
Two self-hosted Generative Artificial Intelligence (GenAI) apps, compared on the facts we track: GitHub popularity and activity, license, language, Docker images and a rough hosting estimate.
AnythingLLM and LLM Harbor are open-source Generative Artificial Intelligence (GenAI) apps you can run on your own server. AnythingLLM is built with Nodejs and licensed under MIT; LLM Harbor is built with Shell under Apache-2.0. On GitHub, AnythingLLM leads with 68.3k stars against 3.3k for LLM Harbor (about 21× as many). Both projects had commits in the last 90 days.
| AnythingLLM | LLM Harbor | |
|---|---|---|
| Description | All-in-one desktop & Docker AI application with built-in RAG, AI agents, No-code agent builder, MCP compatibility, and more. | Containerized LLM toolkit. Run LLM backends, APIs, frontends, and additional services via a concise CLI. |
| GitHub stars | 68,332 | 3,267 |
| Forks | 7.7k | – |
| Open issues (incl. PRs) | 362 | – |
| Commits (12 months) | 1.1k | 1.3k |
| Last activity | 2026-11-09 | 2026-10-31 |
| Latest release | v1.19.0 (2026-10-26) | v0.5.19 (2026-10-31) |
| Score (0–100) | 92 | 86 |
| Maintenance health | Active | Active |
| License | MIT | Apache-2.0 |
| Language / platform | Nodejs, Docker | Docker, Shell |
| Docker Hub image | – | – |
| Docker pulls | – | – |
| Est. server size | ~8 GB, 4 vCPU | ~8 GB, 4 vCPU |
| Est. hosting from | $40/mo | $40/mo |
| Categories | Generative Artificial Intelligence (GenAI) | Generative Artificial Intelligence (GenAI) |
| Links | Website · Source | Website · Source |
Summary: AnythingLLM leads on GitHub popularity. The right choice depends on the features you need — read both project descriptions and docs.
This summary is generated from GitHub, Docker Hub and awesome-selfhosted data as of 2026-11-16. It does not rate features or quality.
Each app gets a score out of 100, computed only from public data we refresh daily:
The score measures community size, upkeep and ease of deployment, not features or quality. Hover a score to see its parts.
Either app fits on a ~8 GB VPS by our rough estimate. Cheapest matching plans:
| Provider | Plan size | ~ USD / month | |
|---|---|---|---|
| Vultr Low-cost cloud compute in 30+ regions. | 8 GB | $40 | Deploy → |
| DigitalOcean Simple droplets, great docs, 1-click Docker image. | 8 GB | $48 | Deploy → |
| Akamai (Linode) Long-running VPS provider with predictable pricing. | 8 GB | $48 | Deploy → |
Prefer managed hosting? Check PikaPods or Elestio.
Estimate from each app's stack and category, not a benchmark. Some hosting links may be affiliate links. We may earn a commission at no extra cost to you; it never affects which apps we list.
AnythingLLM has more GitHub stars (68,332 vs 3,267 for LLM Harbor, as of 2026-11-16). Stars measure interest, not quality.
Commits in the last 12 months: AnythingLLM 1,052, LLM Harbor 1,315. The last recorded activity was 2026-11-09 for AnythingLLM and 2026-10-31 for LLM Harbor. Releases in the past year: AnythingLLM 10+, LLM Harbor 10+. Our health labels: AnythingLLM is “Active”, LLM Harbor is “Active”.
AnythingLLM: MIT. LLM Harbor: Apache-2.0.
We did not detect a Docker Hub image for either app. The upstream listing mentions Docker as a deployment option for AnythingLLM and LLM Harbor. Check each project's documentation for install options.
All-in-one desktop & Docker AI application with built-in RAG, AI agents, No-code agent builder, MCP compatibility, and more.
Containerized LLM toolkit. Run LLM backends, APIs, frontends, and additional services via a concise CLI.
Browse all self-hosted Generative Artificial Intelligence (GenAI) apps, or see every LM Studio alternative.