Pithelo

Five agent repos gaining momentum. Every one measured.

Ranked by stars gained over the last 7 days, growth rate and recent commit activity.

Short candidate pool today: 1 of 5 picks were filled with relaxed rules (1 by star gains since the last snapshot).

Rank 1

crewAIInc/crewAI

CrewAI is an open-source Python framework for building multi-agent workflows, offering Crews (role-based agents that collaborate autonomously) and Flows (event-driven automations with precise control that can also call Crews or make single LLM calls). The same company sells a commercial AMP Suite adding managed deployment, observability, and governance.

Developers building multi-agent systems who want to choose between autonomous role-based collaboration and explicit event-driven control within one framework should look at it.

  • ★ 60,047
  • +★ 268 in 7 days
  • 0.4% growth
  • MIT
  • README ≈ 12,454 tokens to load into context

AI-written summary

Score86.7
Rank 2

code-yeongyu/oh-my-openagent

OmO is a single `omo` command-line agent that runs on senpi, the author's fork of pi, with built-in features including a memory system, CodeMode, and multi-model workflows triggered by keywords like "mass ulw". It installs a native binary via a shell script or `bun`/`npm`, and includes migration tooling (`omo setup`, `omo doctor`) for users coming from the OpenCode edition or LazyCodex.

Developers using OpenCode or LazyCodex, or anyone looking for a terminal agent with multi-model orchestration and session resume/thinking-level controls, may want to evaluate it.

  • ★ 70,297
  • +★ 179 in 7 days
  • 0.3% growth
  • No licence declared
  • README ≈ 4,222 tokens to load into context

AI-written summary

Score63.3
Rank 3

CopilotKit/CopilotKit

CopilotKit is an SDK for building agent-native applications, providing chat UI, generative UI, shared state, and human-in-the-loop workflows for React, Angular, Vue, React Native, Slack, and Microsoft Teams. Its team also makes the AG-UI Protocol, and an optional CopilotKit Intelligence layer adds persistent threads, user memories, analytics, and automatic learning.

Developers who need to put an existing agent in front of users across web, mobile, or team chat apps, with agent-driven UI and user approval steps, should look at it.

  • ★ 38,197
  • +★ 129 in 7 days
  • 0.3% growth
  • MIT
  • README ≈ 5,151 tokens to load into context

AI-written summary

Score50.0
Rank 4

microsoft/autogen

AutoGen is a Python framework for building multi-agent AI applications that can act autonomously or alongside humans, with packages for agent chat, model clients (e.g., OpenAI), MCP tool integration, and a no-code GUI (AutoGen Studio). The repository is archived: Microsoft states it no longer provides bug fixes, security fixes, new features, or support.

Developers with existing AutoGen code should look at the migration guide to Microsoft Agent Framework, and new users should start with that successor instead, since AutoGen is archived and advised against for production use.

  • ★ 61,546
  • +★ 106 in 7 days
  • 0.2% growth
  • CC-BY-4.0
  • README ≈ 6,922 tokens to load into context

AI-written summary

Score0.0
Rank 5

coreyhaines31/marketingskills

Marketing skills for Claude Code and AI agents. CRO, copywriting, SEO, analytics, and growth engineering.

  • ★ 56,256
  • MIT
  • README ≈ 9,071 tokens to load into context

AI-written summary

Score55.0