GitHub: the most-starred AI repositories, read on 3 October
On 3 October 2026, the GitHub API, sorted by stars across four AI topics, places ECC at 271,617, Hermes Agent at 250,839, TensorFlow at 200,674. Stars do not measure quality. Prompt collections are not software.

Original AI-generated illustration · SecuFocus
At a glance
Key points
A snapshot, not a podium. We did not install these repositories.
The ranking
Four searches on 3 October 2026: topics llm, artificial-intelligence, machine-learning and generative-ai, sorted by stars. Among software, the highest counts we read are affaan-m/ECC, 271,617; NousResearch/hermes-agent, 250,839; tensorflow/tensorflow, 200,674; firecrawl/firecrawl, 188,062.
Behind them: AutoGPT, 187,646; ollama/ollama, 182,080; huggingface/transformers, 166,911; langgenius/dify, 157,748. f/prompts.chat shows 171,891 stars. That is a prompt collection, not software. It stays off this list.
Scroll the table sideways to read every column.
| Rank | Repository | Stars | In one line |
|---|---|---|---|
| 1 | ECC | 271,617 | Agent |
| 2 | Hermes Agent | 250,839 | Agent |
| 3 | TensorFlow | 200,674 | Framework |
| 4 | Firecrawl | 188,062 | Web data |
| 5 | AutoGPT | 187,646 | Agent |
| 6 | Ollama | 182,080 | Local models |
| 7 | Transformers | 166,911 | Models |
| 8 | Dify | 157,748 | Workflows |
What leaves
Cloning a repository does not send your files away. Running the tool can. Ollama and Hermes run on your machine, then call a model that may be remote. AutoGPT and Firecrawl talk to the network. A star count does not say what the program sends.
What we did not check
We did not clone these repositories for this note, and we did not review the code. Stars move. This search does not cover all of GitHub: only the first page of each topic. A heavily starred repository without these topics does not appear here.
The choice
The Top Git badge marks popularity read that day, not install advice. If you want a tool, read what it runs and where the text goes. Stars come after.
Check and explore
Sources for this article
Numbers connect each reference to the passages that use it. Dates show when the documentation was consulted.
This article draws on the sources above. The exercises are for you to try on your devices; SecuFocus does not present them as tests carried out by its editorial team. Interfaces and features can change. Method and corrections.
Cite this article
Keep this reference with the article when you save or share it.