This Week's GitHub Stars Ranking

🤖 Agents 2026-11-09 7 min read

Hermes Agent gained 2,264 stars in a week, ranking first in both total and growth; PR Agent gained 1,733 stars in a week, a 12.5% increase, jumping from 87th place overall to second in growth. Across the entire pool, the comparable net increase was 21,531 stars, up 0.374%, and the names and order of the top ten remained unchanged for a week—the first time since the October 19 ranking.

📜 Table of Contents

The top ten didn't change personnel this week, and didn't change positions either — all the movement happened outside the leaderboard. Hermes Agent gained 2,264 stars in a week for a total of 261,759, taking both first places; PR Agent gained 1,733 stars in a week with a total of only 15,598, ranking 87th overall, yet squeezing into second on the growth chart; Lobe Chat gained 747 stars in a week, four times last week's figure. On one side the leaders held steady, on the other small and mid-sized projects got concentrated attention in a short span — those are the two clearest directions on this week's chart.

Data methodology

Data comes from the official GitHub API, with a snapshot time of November 8, 2026 (UTC), compared against a snapshot from the same methodology seven days earlier. The tracking pool has 116 entries, all of which returned values in both weeks — none added, none dropped out. Two of them, Embedchain and Mem0, point to the same repository (a redirect after a rename), with completely duplicated star counts; one was removed from the totals, leaving 115 actually comparable entries. Weekly growth equals current stars minus stars seven days ago, with no smoothing and no adjustment for weekend effects.

After removing the duplicate, the pool grew from 5.7507 million to 5.7723 million, a net gain of 21,531 stars, or 0.374%. Last week, using the same algorithm, it was 19,210 stars and 0.358% — this week is 12.1% faster. One methodological point needs to be spelled out: Hermes Agent only entered the pool on October 30, so last week it had no seven-days-ago comparison figure, and the 2,264 stars were not counted in last week's net gain. Strip that out, and the other 114 entries gained 19,267 stars net this week, essentially flat against last week's 19,210 — the pool's real speed is moving sideways, with no obvious acceleration.

This week's TOP10

RankProjectStarsWeekly growth
1Hermes Agent261,759+2,264
2n8n209,850+551
3Ollama186,382+789
4Awesome Prompts176,545+1,020
5Transformers168,666+248
6Dify160,687+515
7Open WebUI158,136+655
8Langflow156,447+217
9Claude Code155,000+833
10LangChain149,217+378

The top-ten threshold rose from 148,839 to 149,217, up just 378 stars in a week — last week that number was 5,914. The threshold stalled because of the tenth-place project itself: LangChain gained only 378 stars this week, so it is the threshold.

Hermes Agent: no breakout day, but rising all seven days

Spread across seven days, the 2,264 stars work out to daily gains between 293 and 410, averaging 323 a day, with no spike on any single day. That curve looks nothing like a jump from a single trending moment — it looks more like being continuously recommended and continuously starred. By total, it leads second-place n8n by 51,909 stars. By growth, it accounts for 10.5% of the pool's net gain, the single largest chunk from any one project.

It has only been in the pool for ten days, and its earlier baseline was backfilled, so it wasn't on last week's growth chart. This week it took first place on its first appearance in the rankings, beating second-place PR Agent's 1,733 stars by 531. Next week its comparison base becomes a full week, and that growth figure will be one that can be compared horizontally.

PR Agent: up 12.5% in a week, half the gain packed into two days

PR Agent gained 1,733 stars this week for a total of 15,598, a weekly gain of 12.5% — the only double-digit percentage project on the top-ten growth chart, with the other nine all below 1%. It ranks only 87th on the overall chart and can't make any total-count leaderboard, so second on the growth chart is the highest position it can reach.

Breaking out the daily line: of those 1,733 stars, November 2 alone added 470 and November 4 added 403 — 873 combined, or 50.4% of the week. The other five days averaged about 172 stars a day, close to its pace last week. That distribution shows the gain came from one or two concentrated bursts of exposure; the repository itself was renamed and moved under a new organization over these two weeks, and its description even includes a line explicitly stating it is not a free version of Qodo. Whether exposure converts into retention depends on how much daily average it has left next week.

Coding agents slow down across the board, with only Claude Code bucking the trend

Codex CLI gained 1,865 stars last week to take first; this week it's 1,078, down 42%, dropping to fourth. Its daily line is even clearer: 688 stars on October 31 alone, then falling back to between 109 and 275 — last week's momentum is spent. Total 133,816 stars, still twelfth, 15,401 short of the top ten.

Others on the same line are mixed. Claude Code added 833, slightly above last week's 815, one of the few in the top ten to accelerate; OpenHands added 532, versus 619 last week; Cline added 289, Aider 142, Continue 36. The coding-tools layer isn't cooling as a whole — what fell was mainly the one that surged hardest last week.

ComfyUI drops to third, tenth place starts to loosen

ComfyUI gained 1,312 stars this week, 16% below last week's 1,556, dropping from second to third on growth, with a total of 145,793 at 11th overall. Its gap to tenth-place LangChain narrowed from 4,358 stars last week to 3,424. At this week's pace — ComfyUI +1,312, LangChain +378 — it can close 934 stars a week, so that position could change hands within four weeks.

One notch up is narrowing too. Claude Code at 155,000 is just 1,447 behind eighth-place Langflow's 156,447, and with this week's speeds at 833 versus 217, that gap closes in two and a half weeks. Between sixth and seventh, though, little moved: Dify at 160,687 versus Open WebUI at 158,136, a gap of 2,551 that shrank by only 140 in a week.

Who's falling

Sixteen repositories lost stars this week, 97 in total, with another 5 at zero growth. The biggest drops were ChatGLM down 30, GPT Engineer down 28, Chatbot UI down 7, Mistral Inference down 6, and the other twelve down 1 to 4 each. The zero-growth ones are Camunda, Mistral Finetune, Cody, Llama Agents, and AgentVerse — all long-dormant or index-only repositories.

At this scale, on repositories with tens of thousands of stars, that's routine unstarring and doesn't indicate users leaving. More telling is the slow-growth group: Langflow +217, a 0.14% gain, the lowest in the top ten; Transformers +248, 0.15%; and the domestic model repositories ChatGLM, Yi, and InternLM continue to hover near zero growth. Attention on model repositories has been ceded to the tooling layer.

Growth is spreading to small and mid-sized projects

The top ten by growth total 11,112 stars, 51.6% of the pool's net gain; last week the same ten projects took 56.7%. The leaders are absorbing a slightly smaller share, and the emerging names are mid-tier projects like Lobe Chat at 747, llama.cpp at 681, and Open WebUI at 655. Lobe Chat deserves a separate look: it added only 186 last week and 747 this week, with 177 and 180 on November 4 and 5 back to back, for a total of 84,423 at 22nd — the fastest riser from the middle of the chart this week.

One same-day anomaly is worth noting: Anthropic Quickstarts added 250 stars on the final day alone, out of 317 for the week — 80% concentrated in that one day. Total 19,394, a small base, so single-day jumps get amplified easily; next week we'll see whether it has connected to sustained growth or just bounced once.

How to use this chart

Stars reflect attention only, with no direct relationship to quality or performance. Use totals to judge ecosystem maturity and weekly growth to judge where recent attention is flowing — look at both, and don't treat either as an acceptance criterion. For actual selection, release cadence, issue response speed, and running it in your own scenario are more useful than star counts. The direction this week's numbers give: the top ten are locked in position with overall speed flat; growth is spreading to specific tools like PR review and chat interfaces; image workflows and terminal coding agents are both slowing; model repositories stay cold.

Same time next week

The chart updates every Monday, with a fixed data source and methodology for easy week-over-week comparison. If you want to keep this table around to read later, bookmark this page; subscribe to site updates and next week's growth chart and ranking changes will reach you first. If a project you're using rose unusually this week, drop a note in the comments and we'll pull it out for a closer look next time.

## Related articles - [Last week's GitHub Stars ranking: Codex leads with 1,865 stars](/post/github-ai-agent-stars-weekly-20261102) - [The week before's GitHub Stars ranking: ComfyUI gains 4,342 stars in a week](/post/github-ai-agent-stars-weekly-20261026) - [October 19 GitHub Stars ranking: Claude Code adds 1,850 stars weekly](/post/github-ai-agent-stars-weekly-20261019) - [2026 AI Agent Project GitHub Stars Ranking](/post/ai-agent-github-stars-ranking-2026)
Related Articles
2026-08-02
Computer Use Agent Guide 2026: 5 Open Source Ways to Let AI Operate Your Desktop
2026-07-26
AI Workflow vs AI Agent: Key Differences
2026-08-01
MCP Server Example in Python: Build Your First Model Context Protocol Server in 30 Minutes
2026-08-11
AI Agent Examples 2026: 7 Real Working Agents You Can Build This Weekend
2026-07-16
AI Agent Frameworks 2026 Complete Guide
2026-07-17
AI Agent Cost Tracking 2026

Written by our editorial team; tools listed here are tested or verified against public sources. Links point to official sites or GitHub repos for reference only — no paid placements.

💬 Comments (0)

No comments yet. Be the first!

Login to comment