Footer

Your go-to destination for discovering trending open-source projects and uncovering the insights that matter

Legal

  • Terms and conditions
  • Privacy policy

More

  • Pricing
  • Advertise
  • onedump
  • gosnakego
  • Contact us

Social

  • X (Twitter)
  • GitHub
Trendshift owl
What they build speaks louder than words
© 2026 Trendshift, created by Julian Li with ❤️
SignalAdvertise
Trendshift
SignalAdvertise
Sign in
Trendshift owl

Discover trends that matter

SignalAdvertise
DailyWeeklyMonthlyYearly
Live mentionsTopics
RepositoriesDevelopers
InsightsStats

Bifrost
Bifrost

Human-Agent-Society/reef

Infrastructure for continually self‑improving agents

Visit GitHub
Python
7.8k
777
13 contributors
Apache License 2.0
website
AchievementsView all →
FeaturedAdvertise to 180K+ monthly visitors
Bifrost

Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

Momentum

How fast Human-Agent-Society/reef is gaining stars, and where it lands on Trendshift and GitHub Trending

Activity and trending

Stars, forks and merged pull requests each day or month, lined up with the repository's rank on Trendshift Trending

Sep 6 – Nov 5, 2026

Stars +7,390 in 61 days

05001K
616

Forks +739 in 61 days

050100
76

Merged PRs 509 in 61 days

0102030
29

Trendshift Trendingwhere the repository ranked each day it made the top 25

Top 25 overall

1312151320

Top 25 in Python

2114341517587813
Sep 6Sep 14Sep 22Sep 30Oct 8Oct 16Oct 24Nov 1
The left and right arrow keys step from day to day; Home and End go to the first and the last

Stars, forks, merged pull requests and Trendshift Trending rank, by day
DayStars gainedForks gainedMerged PRsAll languagesPython
Sun, Sep 6, 2026+90+114Not trendingNot trending
Mon, Sep 7, 2026+123+813Not trendingNot trending
Tue, Sep 8, 2026+80+112Not trendingNot trending
Wed, Sep 9, 2026+121+612Not trendingNot trending
Thu, Sep 10, 2026+26+66Not trendingNot trending
Fri, Sep 11, 2026+51+716Not trendingNot trending
Sat, Sep 12, 2026+143+220Not trendingNot trending
Sun, Sep 13, 2026+105+3010Not trendingNot trending
Mon, Sep 14, 2026+265+29Not trending#21
Tue, Sep 15, 2026+489+1011Not trending#14
Wed, Sep 16, 2026+607+768#13#3
Thu, Sep 17, 2026+570+6310#12#4
Fri, Sep 18, 2026+403+405Not trending#15
Sat, Sep 19, 2026+111+113Not trendingNot trending
Sun, Sep 20, 2026+114+1210Not trendingNot trending
Mon, Sep 21, 2026+159+1210Not trendingNot trending
Tue, Sep 22, 2026+127+148Not trendingNot trending
Wed, Sep 23, 2026+222+114Not trendingNot trending
Thu, Sep 24, 2026+367+536Not trending#17
Fri, Sep 25, 2026+501+596#15#5
Sat, Sep 26, 2026+616+570#13#8
Sun, Sep 27, 2026+560+6010#20#7
Mon, Sep 28, 2026+546+659Not trending#8
Tue, Sep 29, 2026+444+539Not trending#13
Wed, Sep 30, 2026+79+49Not trendingNot trending
Thu, Oct 1, 2026+67+37Not trendingNot trending
Fri, Oct 2, 2026+14+42Not trendingNot trending
Sat, Oct 3, 2026+1+10Not trendingNot trending
Sun, Oct 4, 2026+3+12Not trendingNot trending
Mon, Oct 5, 2026+0+24Not trendingNot trending
Tue, Oct 6, 2026+4+26Not trendingNot trending
Wed, Oct 7, 2026+9+06Not trendingNot trending
Thu, Oct 8, 2026+20+23Not trendingNot trending
Fri, Oct 9, 2026+10+22Not trendingNot trending
Sat, Oct 10, 2026+0+13Not trendingNot trending
Sun, Oct 11, 2026+5+10Not trendingNot trending
Mon, Oct 12, 2026+46+515Not trendingNot trending
Tue, Oct 13, 2026+17+212Not trendingNot trending
Wed, Oct 14, 2026+14+217Not trendingNot trending
Thu, Oct 15, 2026+10+25Not trendingNot trending
Fri, Oct 16, 2026+5+03Not trendingNot trending
Sat, Oct 17, 2026+7+06Not trendingNot trending
Sun, Oct 18, 2026+16+45Not trendingNot trending
Mon, Oct 19, 2026+35+67Not trendingNot trending
Tue, Oct 20, 2026+31+712Not trendingNot trending
Wed, Oct 21, 2026+20+23Not trendingNot trending
Thu, Oct 22, 2026+23+36Not trendingNot trending
Fri, Oct 23, 2026+10+316Not trendingNot trending
Sat, Oct 24, 2026+0+010Not trendingNot trending
Sun, Oct 25, 2026+31+100Not trendingNot trending
Mon, Oct 26, 2026+16+25Not trendingNot trending
Tue, Oct 27, 2026+14+15Not trendingNot trending
Wed, Oct 28, 2026+2+317Not trendingNot trending
Thu, Oct 29, 2026+6+222Not trendingNot trending
Fri, Oct 30, 2026+8+012Not trendingNot trending
Sat, Oct 31, 2026+10+03Not trendingNot trending
Sun, Nov 1, 2026+3+010Not trendingNot trending
Mon, Nov 2, 2026+4+129Not trendingNot trending
Tue, Nov 3, 2026+1+216Not trendingNot trending
Wed, Nov 4, 2026+7+08Not trendingNot trending
Thu, Nov 5, 2026+2+00Not trendingNot trending
Total+7,390+739509
  • 1#1 on the list
  • 7Rank that day
  • Not trending
  • Mentions and achievements

    Where people are talking about Human-Agent-Society/reef, and every badge it has earned
    1. Oct 291 post
      • ai agents repos worth checking: 1. Foreman a supervisor for coding agents. checks progress, requirements and test evidence against the original task before deciding whether the worker is done. 2.…

        Read full on X, posted by @5haman

        @5haman · x.com

    2. 21 days earlier
    3. Oct 81 post
      • reef is the first open-source infrastructure for continual self-improving agents. It puts agent inference, feedback, learning and versioned delivery in one loop. If you're building agents that should…

        Read full on X, posted by @Xuan04263449

        @Xuan04263449 · x.com

    4. Oct 61 post
      • (7/8) When is a small guide enough? 8B suffices on Polyomino and Lasso. On more saturated tasks (TriMul, AHC058), 14B helps through finer specialization. Our probes suggest that 14B more precisely…

        Read full on X, posted by @ao_qu18465

        @ao_qu18465 · x.com

    5. 8 days earlier
    6. Sep 282 posts7 posts
      • the "how" that makes it not sketchy: every /reefine change is a candidate. it runs against an eval before touching production. fail → nothing ships. pass → hot-swapped in, old version kept for…

        Read full on X, posted by @heyfredds

        @heyfredds · x.com

      • Reef flips this: the inference service becomes where learning starts. → Records every request → Matches feedback → Builds an update → Tests it first Passes? Hot-swapped, zero downtime. Fails? Nothing…

        Read full on X, posted by @marcusyul

        @marcusyul · x.com

      • this is the part that makes reef a new category. model weights aren't the only thing that updates. the whole harness evolves too prompts, skills, tools, rules, config. both surfaces, same loop, from…

        Read full on X, posted by @liambraus

        @liambraus · x.com

      • One loop connects it all: serving → feedback → learning → evaluation → versioned delivery It can update model weights (GPU) or the agent harness: prompts, skills, tools, rules (no GPU needed). Every…

        Read full on X, posted by @aresotik

        @aresotik · x.com

      • and weights aren't the only thing that evolves. /reefine grows the harness itself prompts, skills, tools, rules, config. describe the capability, reef writes it, evaluates it, ships it. the demo adds…

        Read full on X, posted by @mattknowsai

        @mattknowsai · x.com

      • Every accepted update becomes a new version with full history. → See what changed → See why it changed → Roll back anytime Even rejected candidates stay visible. Self-improvement you can actually…

        Read full on X, posted by @nattcornelis

        @nattcornelis · x.com

      • And you don't have to switch tools. Reef plugs into the harnesses you already use through adapters: Claude Code, Codex, OpenCode, Hermes, Pi and Terminus 2. Each scenario keeps its own data and…

        Read full on X, posted by @zentalksai

        @zentalksai · x.com

    You might also like

    Repositories close to Human-Agent-Society/reef
    FeaturedAdvertise to 180K+ monthly visitors
    Bifrost

    Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

    raia-live/amfs

    Python

    Recursive Self-Improvement for AI agents and models, helping your agents self-improve with every run

    9014

    ReflexioAI/reflexio

    Python

    Make your agents improve themselves. Reflexio is an AI agent self-improvement harness that enables your AI agents to continuously learn from real user interactions.

    38354

    AgentToolkit/altk-evolve

    Python

    Self improving agents through iterations

    13521

    greyhaven-ai/autocontext

    Python

    a recursive self-improving harness designed to help your agents (and future iterations of those agents) succeed on any task

    1.3k127

    overmind-core/overmind

    Python

    The platform for continuously improving AI agents.

    74881

    kayba-ai/agentic-context-engine

    Python

    🧠 Make your agents learn from experience. Now available as a hosted solution at kayba.ai

    2.6k326

    unifyai/unify-agent

    Python

    A self-improving agent harness for research on how an assistant learns tasks and skills

    13416

    kayba-ai/recursive-improve

    Python

    🪞 Make your agents recursively self-improve

    30035

    canvas-org/meta-agent

    Python

    Continual harness optimization

    817