Agentic-Index

🌐 Agentic-Index – The Data-Driven AI-Agent Repository Index

Agentic-Index continuously scores and curates every open-source framework for building autonomous AI agents. Fast search, transparent metrics, zero BS.

We rank everything using a transparent scoring formula based on:

🎯 TL;DR: This isn’t just a listβ€”it’s your launchpad for building with AI agents.

Want a shortcut? Jump to the Fast-Start table.

build ![coverage](https://img.shields.io/badge/coverage-80%25-brightgreen) docs Site license PyPI Release Notes

This catalogue is maintained by the Agentic-Index project and is updated regularly (aiming for monthly refreshes) to reflect the rapidly evolving landscape of Agentic-AI.

TOC

✨ Why Agentic Index is Different

In the fast-moving world of Agentic-AI, finding high-quality, actively maintained, and truly impactful frameworks can be a pain. Many lists are subjective or just track stars. Agentic-Index cuts through the noise with an analytical approach:

Agentic-Index is built to be a reliable, data-driven launchpad for your next Agentic-AI project.

⚑ Installation & Quick-start

pip install agentic-index-cli

agentic-index scrape --min-stars 100
agentic-index enrich data/repos.json
agentic-index rank data/repos.json
cat README.md | less         # see table injected

πŸš€ Fast-Start Picks (Curated for Newcomers)

New to Agentic-AI or just want the good stuff fast? These repos are top-tier for usability, community, docs, or just plain cool ideas:

πŸ† The Agentic-Index Top 50: AI Agent Repositories

The definitive list of Agentic-AI repositories, ranked by the Agentic Index Score. This score is a holistic measure of project quality, activity, and community love. (Data updated as of: {timestamp} UTC)

| Rank | Repo | Score | β–² StarsΞ” | β–² ScoreΞ” | Category | |—–:|β€”β€”|β€”β€”:|β€”β€”-:|——–:|β€”β€”β€”-| | 1 | dify | 6.08 | | | General-purpose | | 2 | langflow | 5.95 | | | DevTools | | 3 | browser-use | 5.88 | | | General-purpose | | 4 | OpenHands | 5.84 | | | General-purpose | | 5 | lobe-chat | 5.83 | | | RAG-centric | | 6 | MetaGPT | 5.82 | | | Multi-Agent Coordination | | 7 | ragflow | 5.81 | | | RAG-centric | | 8 | LLaMA-Factory | 5.79 | | | General-purpose | | 9 | system-prompts-and-models-of-ai-tools | 5.78 | | | DevTools | | 10 | cline | 5.72 | | | General-purpose | | 11 | anything-llm | 5.71 | | | RAG-centric | | 12 | llama_index | 5.68 | | | General-purpose | | 13 | autogen | 5.67 | | | General-purpose | | 14 | awesome-llm-apps | 5.63 | | | RAG-centric | | 15 | Flowise | 5.60 | | | General-purpose | | 16 | mem0 | 5.57 | | | General-purpose | | 17 | ChatTTS | 5.56 | | | General-purpose | | 18 | Langchain-Chatchat | 5.56 | | | RAG-centric | | 19 | crewAI | 5.55 | | | Multi-Agent Coordination | | 20 | AgentGPT | 5.51 | | | General-purpose | | 21 | agno | 5.47 | | | Multi-Agent Coordination | | 22 | khoj | 5.46 | | | Experimental | | 23 | ChatDev | 5.45 | | | Multi-Agent Coordination | | 24 | LibreChat | 5.45 | | | General-purpose | | 25 | ai-agents-for-beginners | 5.43 | | | General-purpose | | 26 | cherry-studio | 5.43 | | | General-purpose | | 27 | Jobs_Applier_AI_Agent_AIHawk | 5.43 | | | General-purpose | | 28 | qlib | 5.42 | | | Experimental | | 29 | composio | 5.37 | | | General-purpose | | 30 | FastGPT | 5.36 | | | RAG-centric | | 31 | gpt-researcher | 5.35 | | | Experimental | | 32 | CopilotKit | 5.33 | | | General-purpose | | 33 | haystack | 5.33 | | | RAG-centric | | 34 | swarm | 5.26 | | | Multi-Agent Coordination | | 35 | agentic | 5.24 | | | General-purpose | | 36 | vanna | 5.23 | | | RAG-centric | | 37 | DB-GPT | 5.21 | | | General-purpose | | 38 | deep-research | 5.21 | | | Experimental | | 39 | letta | 5.21 | | | General-purpose | | 40 | agenticSeek | 5.20 | | | General-purpose | | 41 | SWE-agent | 5.20 | | | General-purpose | | 42 | eliza | 5.19 | | | General-purpose | | 43 | RagaAI-Catalyst | 5.19 | | | RAG-centric | | 44 | DocsGPT | 5.18 | | | DevTools | | 45 | awesome-ai-agents | 5.17 | | | General-purpose | | 46 | devika | 5.14 | | | Experimental | | 47 | goose | 5.14 | | | General-purpose | | 48 | suna | 5.13 | | | General-purpose | | 49 | SuperAGI | 5.13 | | | RAG-centric | | 50 | ai-pdf-chatbot-langchain | 5.12 | | | General-purpose | ➑️ Dig into how these scores are cooked up in our Methodology section and the full recipe in /docs/methodology.md.

πŸ’Ž Honourable Mentions / Niche & Novel Gems

Beyond the top-ranked, these projects are cooking up unique ideas, serving specific niches, or pushing experimental boundaries in Agentic-AI:

Our Methodology & Scoring Explained

<details> <summary>πŸ”¬ Our Methodology & Scoring Explained (Click to Expand)</summary>

Agentic-Index believes in full transparency. Here’s the lowdown on how we find, vet, and score repositories.

The core Agentic-Index Scoring Formula: Score = 0.35*log2(stars+1) + 0.20*recency_factor + 0.15*issue_health + 0.15*doc_completeness + 0.10*license_freedom + 0.05*ecosystem_integration<sup>†</sup>

<sup>†</sup> Weights are reviewed and potentially tuned quarterly. Full math and reasoning in /docs/methodology.md.

Quick Look at Components:

For the full, unabridged version, see ./docs/methodology.md.

</details>

🏷️ Category Definitions

Quick guide to our categories (and the icons you’ll see in the table):

πŸ”„ Changelog

This isn’t a static list. It’s alive! See CHANGELOG.md for all the adds, drops, and major rank shuffles.

πŸ— Architecture

flowchart LR
    A[User] --> B[Scrape]
    B --> C[JSON]
    C --> D[Rank]
    D --> E[Markdown]
    E --> F[View]

πŸ”§ Usage

Run the indexer to fetch fresh repo data:

python -m agentic_index_cli.agentic_index --min-stars 50 --iterations 1 --output data

Generated tables live in the data/ directory.

πŸ”„ How refresh works

A scheduled GitHub Action keeps the index up to date. It runs the scraper and ranker, opens a pull request with any changes, and can auto-merge when all checks pass. You can also trigger this process manually by running scripts/trigger_refresh.sh.

πŸ§ͺ Testing

This project uses pytest for unit tests and pa11y for accessibility checks. Ensure Chrome is installed before running pa11y:

# via puppeteer
npx puppeteer browsers install chrome
# or with apt
sudo apt-get install -y chromium

Run tests with:

pytest -q

CI runs tests with network access disabled. Set CI_OFFLINE=1 or run pytest --disable-socket locally to replicate the offline environment.

To check accessibility after building the site:

npx pa11y web/index.html

You can also run ./scripts/install_pa11y_deps.sh to install pa11y and Chrome.

πŸ’» Developer

To trigger a data refresh via GitHub Actions, run:

bash scripts/trigger_refresh.sh 75

Replace 75 with your desired minimum star count. The script requires the GitHub CLI and an authenticated token.

🀝 How to Contribute

Agentic-Index aims to be the spot for Agentic-AI frameworks. Your brainpower and suggestions are gold.

Check out CONTRIBUTING.md for how to:

For tips on keeping your branch in sync with main and resolving conflicts, see CONFLICT_RESOLUTION.md.

Let’s build the best damn agent list together! Code of Conduct

πŸ›‘ Code of Conduct

Please see our Code of Conduct for contributor expectations.

πŸ“œ License

The content of Agentic-Index (this README.md, files in /docs/, etc.) is licensed under(https://creativecommons.org/licenses/by-sa/4.0/).

Any scripts or code for analysis and generation (e.g., in /scripts, if we add β€˜em) are licensed under(https://opensource.org/licenses/MIT).

Β© 2025 Agentic-Index Maintainers

Last Sync Top Repo