Open-Source AI Statistics 2026
What the open-source AI world actually looks like, counted from primary data: 2,123 projects, each one verified against its GitHub repository, catalogued and categorised by LibreAI.
Last updated: 1 November 2026 · Publisher: LibreAI · By the LibreAI team · Data snapshot: October 2026
Key takeaways
- LibreAI's catalogue holds 2,123 open-source AI projects, every one verified against GitHub.
- The projects span 17 categories; AI Agents is the largest with 483 projects (22.7%).
- Together the 2,123 projects hold 24,275,545 GitHub stars and 3,117,858 forks.
- The median catalogued project has 1,934 stars — the typical entry is established, not a toy.
- 117 projects have 50,000+ stars, and 496 have 10,000 or more.
- 1,126 projects carry a full install guide on LibreAI; 997 are lesser-known "hidden gems" on shorter pages.
- MIT is the most common licence (913 projects, 43.0%), followed by Apache-2.0 (535, 25.2%).
- Permissive licences (MIT, Apache-2.0, BSD, Unlicense) cover 1,468 projects — 69.1% of the catalogue.
- AI Agents projects also hold the most stars of any category: 8,749,971 (36.0% of all stars).
- The 2,123 projects come from 1,890 distinct GitHub publishers.
- 1,985 projects (93.5%) maintain their own website beyond their GitHub page.
- The most-starred project in the catalogue is openclaw, with 391,625 stars.
Source: LibreAI catalogue data, verified against GitHub (snapshot: October 2026) — libreaihub.com
How many open-source AI projects are there?
LibreAI has verified 2,123 open-source AI projects against GitHub. Every project in the count has a live GitHub repository that was checked for its stars, forks, licence and last-push date, and a unique page in the catalogue — the count contains no duplicates (2,123 unique projects from 1,890 distinct publishers).
Source: LibreAI catalogue data, verified against GitHub (snapshot: October 2026) — libreaihub.com
Which category of open-source AI is the biggest?
AI Agents is the biggest category, with 483 projects — more than Dev Tools (254) and Audio (151) combined. The full count:
Projects per category, top 10 of 17 · Chart: LibreAI
| Category | Projects | Combined stars |
|---|---|---|
| AI Agents | 483 | 8,749,971 |
| Dev Tools | 254 | 3,771,013 |
| Audio | 151 | 1,409,037 |
| Video | 144 | 1,105,500 |
| Local LLM | 136 | 1,996,762 |
| Image | 134 | 1,397,158 |
| Automation | 128 | 1,253,346 |
| RAG & Knowledge | 92 | 926,768 |
| Self-Hosted | 81 | 1,287,541 |
| AI 3D | 77 | 207,837 |
| Easy Install | 75 | 315,338 |
| AI Music | 75 | 50,904 |
| Clipping | 70 | 145,693 |
| Design | 63 | 727,387 |
| Writing | 61 | 247,023 |
| Productivity | 51 | 517,960 |
| AI Chat | 48 | 166,307 |
Source: LibreAI catalogue data, verified against GitHub (snapshot: October 2026) — libreaihub.com
How popular are open-source AI projects on GitHub?
Very — the catalogue's projects hold 24,275,545 stars and 3,117,858 forks between them. Popularity is top-heavy: the mean project has 11,435 stars but the median has 1,934, because a small group of very large projects pulls the average up.
| GitHub stars | Projects | Share |
|---|---|---|
| <100 stars | 282 | 13.3% |
| 100–999 stars | 564 | 26.6% |
| 1,000–9,999 stars | 781 | 36.8% |
| 10,000–49,999 stars | 379 | 17.9% |
| 50,000+ stars | 117 | 5.5% |
Source: LibreAI catalogue data, verified against GitHub (snapshot: October 2026) — libreaihub.com
Which open-source AI project has the most GitHub stars?
openclaw has the most stars in the catalogue, with 391,625. The ten most-starred projects:
| Project | Stars | Forks | Category | Licence |
|---|---|---|---|---|
| openclaw | 391,625 | 82,288 | AI Agents | MIT |
| awesome-selfhosted | 323,662 | 15,154 | Self-Hosted | Unknown |
| superpowers | 294,899 | 26,354 | AI Agents | MIT |
| ECC | 275,200 | 41,075 | AI Agents | MIT |
| hermes-agent | 252,044 | 54,288 | AI Agents | MIT |
| deepseek-harness | 245,484 | 29,409 | AI Agents | MIT |
| OpenCode | 211,612 | 28,116 | Dev Tools | MIT |
| n8n | 206,576 | 60,984 | Automation | Custom (check repo) |
| yt-dlp | 195,267 | 16,961 | Video | Unlicense |
| firecrawl | 189,612 | 10,047 | AI Agents | AGPL-3.0 |
Source: LibreAI catalogue data, verified against GitHub (snapshot: October 2026) — libreaihub.com
What licences do open-source AI projects use?
MIT is the licence of choice: 913 projects (43.0%) use it, and permissive licences overall cover 69.1% of the catalogue. Copyleft licences (AGPL and GPL family) account for 265 projects (12.5%). Licence names below are GitHub's own detection results, quoted exactly as recorded — "Custom", "NOASSERTION" and "Unknown" mean GitHub could not match a standard licence text.
| Licence | Projects | Share |
|---|---|---|
| MIT | 913 | 43.0% |
| Apache-2.0 | 535 | 25.2% |
| Custom | 168 | 7.9% |
| AGPL-3.0 | 149 | 7.0% |
| GPL-3.0 | 102 | 4.8% |
| NOASSERTION | 99 | 4.7% |
| Unknown | 62 | 2.9% |
| Custom (check repo) | 26 | 1.2% |
| BSD-3-Clause | 12 | 0.6% |
| GPL-2.0 | 11 | 0.5% |
Source: LibreAI catalogue data, verified against GitHub (snapshot: October 2026) — libreaihub.com
How many projects have a full install guide?
1,126 of the 2,123 projects have a full install guide on LibreAI — step-by-step setup, features, use cases and FAQ. The remaining 997 are catalogued as "hidden gems": smaller or newer projects on shorter pages that few other sites have covered. All 2,123 project pages include the project's key facts, and 1,495 of them name the main alternatives a reader should compare.
Source: LibreAI catalogue data, verified against GitHub (snapshot: October 2026) — libreaihub.com
How active are these projects?
Active. In the catalogue's October 2026 verification snapshot, 2,048 of the 2,123 projects (96.5%) recorded their most recent GitHub push in 2026, and the freshest push in the data is dated 26 October 2026. Only 75 projects show a last-push date before 2026 — recorded as 46 in 2025, 23 in 2024, 5 in 2023 and 1 in 2022.
Source: LibreAI catalogue data, verified against GitHub (snapshot: October 2026) — libreaihub.com
Methodology and sources
Every statistic on this page comes from one primary source: the LibreAI catalogue itself. Each listed project was verified against its GitHub repository, and the figures published here — project counts, category counts, stars, forks, licences and last-push dates — were counted directly from that verified data in November 2026 (data snapshot: October 2026). No figure is copied from another statistics roundup, and no figure describes projects outside this catalogue: this page measures the 2,123 projects LibreAI lists, not the whole of GitHub. Star and fork counts move daily on GitHub; the numbers here are the recorded snapshot and are refreshed quarterly. The catalogue does not record programming languages, so no language statistics are published.
LibreAI (2026). Open-Source AI Statistics 2026. https://libreaihub.com/open-source-ai-statistics/
Questions about the data, or a project that should be listed? See the contact page.