GitHub AI Agent Repository Stars Tracker
Weekly tracking of top 30 AI Agent repositories on GitHub by star count. TypeScript and Python dominate with 70% of ranked projects, while Claude Code ecosystem shows rapid growth.
Data Overview
- Last Updated: 2026-03-13
- Update Frequency: Weekly
- Primary Sources: GitHub API - AI Agent Topic Search
This tracker monitors the top 30 AI Agent repositories on GitHub ranked by star count. The dataset captures ecosystem dynamics, language preferences, and emerging projects in the rapidly evolving AI agent landscape.
Methodology
Data Collection:
- Source: GitHub Search API with topic filter
ai-agent - Filter: Repositories with >1000 stars
- Sort: Descending by star count
- Limit: Top 30 repositories
Metrics Captured:
- Stars: Primary popularity metric
- Forks: Community engagement indicator
- Language: Primary programming language
- Description: Project purpose (summarized)
Update Schedule: Weekly snapshots every Thursday at 00:00 UTC
Inclusion Criteria:
- Repository must have
ai-agenttopic tag - Minimum 1000 stars at collection time
- Public visibility only
Current Data
| Rank | Repository | Stars | Forks | Language | Description |
|---|---|---|---|---|---|
| 1 | zhayujie/chatgpt-on-wechat | 42,171 | 9,819 | Python | CowAgent - AI assistant with task planning, OS access, and long-term memory |
| 2 | CherryHQ/cherry-studio | 41,382 | 3,824 | TypeScript | AI productivity studio with 300+ assistants and unified LLM access |
| 3 | CopilotKit/CopilotKit | 29,325 | 3,815 | TypeScript | Frontend Stack for Agents & Generative UI, AG-UI Protocol creator |
| 4 | shareAI-lab/learn-claude-code | 25,751 | 4,623 | TypeScript | Nano Claude Code-like agent built from scratch |
| 5 | activepieces/activepieces | 21,196 | 3,409 | TypeScript | AI Agents & MCPs with ~400 MCP servers for AI agents |
| 6 | googleworkspace/cli | 19,711 | 883 | Rust | Google Workspace CLI with AI agent skills for Drive, Gmail, Calendar |
| 7 | iOfficeAI/AionUi | 18,633 | 1,455 | TypeScript | Free local cowork app with OpenClaw for multiple AI CLIs |
| 8 | trycua/cua | 13,042 | 805 | Python | Computer-Use Agents infrastructure with sandboxes and SDKs |
| 9 | e2b-dev/E2B | 11,261 | 802 | MDX | Secure environment with real-world tools for enterprise agents |
| 10 | saturndec/waoowaoo | 9,272 | 2,013 | TypeScript | Professional AI Agent platform for film & video production |
| 11 | Panniantong/Agent-Reach | 8,823 | 634 | Python | AI agent with internet visibility for Twitter, Reddit, YouTube |
| 12 | frankbria/ralph-claude-code | 7,779 | 556 | Shell | Autonomous AI development loop for Claude Code |
| 13 | alibaba/OpenSandbox | 7,676 | 565 | Python | Sandbox platform for AI with multi-language SDKs |
| 14 | crestalnetwork/intentkit | 6,488 | 700 | Python | Self-hosted cloud agent cluster for collaborative AI teams |
| 15 | NousResearch/hermes-agent | 6,409 | 710 | Python | Agent that grows with users |
| 16 | UfoMiao/zcf | 5,668 | 395 | TypeScript | Zero-Config Code Flow for Claude Code & Codex |
| 17 | mahseema/awesome-ai-tools | 4,574 | 1,096 | Other | Curated list of AI tools |
| 18 | Integuru-AI/Integuru | 4,555 | 360 | Python | AI agent building permissionless integrations via API reverse engineering |
| 19 | presenton/presenton | 4,324 | 838 | HTML | Open-source AI presentation generator |
| 20 | FedML-AI/FedML | 4,010 | 764 | Python | Distributed training, model serving, and federated learning |
| 21 | UnicomAI/wanwu | 3,919 | 79 | Go | China Unicom enterprise-grade multi-tenant agent platform |
| 22 | groupultra/telegram-search | 3,820 | 243 | TypeScript | Telegram chat history export and fuzzy search |
| 23 | LazyAGI/LazyLLM | 3,747 | 364 | Python | Multi-agent LLM application framework |
| 24 | zebbern/claude-code-guide | 3,588 | 333 | Other | Claude Code setup guide with commands, workflows, and tips |
| 25 | mindfold-ai/Trellis | 3,576 | 181 | Python | All-in-one AI framework & toolkit |
| 26 | filipecalegario/awesome-vibe-coding | 3,552 | 368 | Other | Curated list for AI-assisted coding references |
| 27 | smallcloudai/refact | 3,519 | 307 | Rust | AI Agent for end-to-end engineering tasks |
| 28 | SWE-agent/mini-swe-agent | 3,291 | 455 | Python | 100-line AI agent scoring >74% on SWE-bench verified |
| 29 | Intelligent-Internet/ii-agent | 3,175 | 488 | Python | Open-source framework for intelligent agents |
| 30 | i-am-bee/beeai-framework | 3,161 | 418 | Python | Production-ready AI agents in Python and TypeScript |
Statistics Summary
| Metric | Value |
|---|---|
| Total Stars (Top 30) | 302,421 |
| Total Forks (Top 30) | 36,114 |
| Average Stars | 10,081 |
| Median Stars | 4,299 |
| Total Repos with >1000 stars | 64 |
Language Distribution
| Language | Count | Percentage |
|---|---|---|
| TypeScript | 11 | 36.7% |
| Python | 10 | 33.3% |
| Rust | 2 | 6.7% |
| Shell | 1 | 3.3% |
| MDX | 1 | 3.3% |
| HTML | 1 | 3.3% |
| Go | 1 | 3.3% |
| Other/Markdown | 4 | 13.3% |
Trends & Observations
Ecosystem Concentration: The top 3 repositories account for 37% of total stars among the top 30, indicating significant concentration at the ecosystemβs apex. chatgpt-on-wechat leads with 42K stars, followed closely by cherry-studio at 41K.
Language Landscape: TypeScript and Python together represent 70% of ranked projects. TypeScript leads narrowly at 36.7% (11 repos), with Python at 33.3% (10 repos). This split reflects the dual nature of AI agent development: Python for ML/AI capabilities, TypeScript for production interfaces and tooling.
Claude Code Ecosystem Emergence: Four repositories in the top 30 directly relate to Claude Code tooling:
- learn-claude-code (#4): 25,751 stars
- AionUi (#7): OpenClaw for Claude Code support
- ralph-claude-code (#12): 7,779 stars
- claude-code-guide (#24): 3,588 stars
Combined, Claude Code-related projects hold approximately 40,000 stars, representing 13% of the top 30 total.
Geographic Diversity: Chinese-developed projects hold two top-5 positions (chatgpt-on-wechat, cherry-studio), plus China Unicomβs enterprise platform at #21. This suggests strong parallel development ecosystems in both Western and Chinese markets.
Specialization Patterns:
- Frontend/Tooling: CopilotKit, activepieces, AionUi
- Sandboxes/Runtime: E2B, OpenSandbox, CUA
- Frameworks: LazyLLM, beeai-framework, ii-agent
- Productivity: cherry-studio, presenton
Fork-to-Star Ratios: chatgpt-on-wechat shows the highest engagement with 9,819 forks (23% of stars), suggesting strong community contribution. Projects with fork ratios below 10% (AionUi, wanwu) indicate more consumption-focused user bases.
πΊ Scout Intel: What Others Missed
Confidence: high | Novelty Score: 50/100
While star counts dominate discussions of repository popularity, the fork-to-star ratio reveals a more nuanced picture of ecosystem health. chatgpt-on-wechat maintains a 23% fork ratio, suggesting active contributors building extensions. In contrast, cherry-studio at 9.2% and AionUi at 7.8% indicate consumption patterns rather than contribution. The emergence of Claude Code as a distinct category with 4 top-30 repos signals a shift from general-purpose agent frameworks toward tooling for specific AI coding assistants. With 40,000 combined stars, Claude Code projects now rival established categories like sandbox infrastructure (E2B at 11K) and computer-use agents (CUA at 13K).
Key Implication: Builders targeting the AI agent market should prioritize TypeScript for user-facing tooling and Python for core agent logic. The Claude Code ecosystem represents an emerging opportunity with lower competition than the saturated general-purpose framework space.
Changelog
| Date | Change | Details |
|---|---|---|
| 2026-03-13 | added | Initial data collection - baseline snapshot of top 30 AI Agent repositories |
Sources
- GitHub API - AI Agent Repositories β GitHub, 2026-03-13
GitHub AI Agent Repository Stars Tracker
Weekly tracking of top 30 AI Agent repositories on GitHub by star count. TypeScript and Python dominate with 70% of ranked projects, while Claude Code ecosystem shows rapid growth.
Data Overview
- Last Updated: 2026-03-13
- Update Frequency: Weekly
- Primary Sources: GitHub API - AI Agent Topic Search
This tracker monitors the top 30 AI Agent repositories on GitHub ranked by star count. The dataset captures ecosystem dynamics, language preferences, and emerging projects in the rapidly evolving AI agent landscape.
Methodology
Data Collection:
- Source: GitHub Search API with topic filter
ai-agent - Filter: Repositories with >1000 stars
- Sort: Descending by star count
- Limit: Top 30 repositories
Metrics Captured:
- Stars: Primary popularity metric
- Forks: Community engagement indicator
- Language: Primary programming language
- Description: Project purpose (summarized)
Update Schedule: Weekly snapshots every Thursday at 00:00 UTC
Inclusion Criteria:
- Repository must have
ai-agenttopic tag - Minimum 1000 stars at collection time
- Public visibility only
Current Data
| Rank | Repository | Stars | Forks | Language | Description |
|---|---|---|---|---|---|
| 1 | zhayujie/chatgpt-on-wechat | 42,171 | 9,819 | Python | CowAgent - AI assistant with task planning, OS access, and long-term memory |
| 2 | CherryHQ/cherry-studio | 41,382 | 3,824 | TypeScript | AI productivity studio with 300+ assistants and unified LLM access |
| 3 | CopilotKit/CopilotKit | 29,325 | 3,815 | TypeScript | Frontend Stack for Agents & Generative UI, AG-UI Protocol creator |
| 4 | shareAI-lab/learn-claude-code | 25,751 | 4,623 | TypeScript | Nano Claude Code-like agent built from scratch |
| 5 | activepieces/activepieces | 21,196 | 3,409 | TypeScript | AI Agents & MCPs with ~400 MCP servers for AI agents |
| 6 | googleworkspace/cli | 19,711 | 883 | Rust | Google Workspace CLI with AI agent skills for Drive, Gmail, Calendar |
| 7 | iOfficeAI/AionUi | 18,633 | 1,455 | TypeScript | Free local cowork app with OpenClaw for multiple AI CLIs |
| 8 | trycua/cua | 13,042 | 805 | Python | Computer-Use Agents infrastructure with sandboxes and SDKs |
| 9 | e2b-dev/E2B | 11,261 | 802 | MDX | Secure environment with real-world tools for enterprise agents |
| 10 | saturndec/waoowaoo | 9,272 | 2,013 | TypeScript | Professional AI Agent platform for film & video production |
| 11 | Panniantong/Agent-Reach | 8,823 | 634 | Python | AI agent with internet visibility for Twitter, Reddit, YouTube |
| 12 | frankbria/ralph-claude-code | 7,779 | 556 | Shell | Autonomous AI development loop for Claude Code |
| 13 | alibaba/OpenSandbox | 7,676 | 565 | Python | Sandbox platform for AI with multi-language SDKs |
| 14 | crestalnetwork/intentkit | 6,488 | 700 | Python | Self-hosted cloud agent cluster for collaborative AI teams |
| 15 | NousResearch/hermes-agent | 6,409 | 710 | Python | Agent that grows with users |
| 16 | UfoMiao/zcf | 5,668 | 395 | TypeScript | Zero-Config Code Flow for Claude Code & Codex |
| 17 | mahseema/awesome-ai-tools | 4,574 | 1,096 | Other | Curated list of AI tools |
| 18 | Integuru-AI/Integuru | 4,555 | 360 | Python | AI agent building permissionless integrations via API reverse engineering |
| 19 | presenton/presenton | 4,324 | 838 | HTML | Open-source AI presentation generator |
| 20 | FedML-AI/FedML | 4,010 | 764 | Python | Distributed training, model serving, and federated learning |
| 21 | UnicomAI/wanwu | 3,919 | 79 | Go | China Unicom enterprise-grade multi-tenant agent platform |
| 22 | groupultra/telegram-search | 3,820 | 243 | TypeScript | Telegram chat history export and fuzzy search |
| 23 | LazyAGI/LazyLLM | 3,747 | 364 | Python | Multi-agent LLM application framework |
| 24 | zebbern/claude-code-guide | 3,588 | 333 | Other | Claude Code setup guide with commands, workflows, and tips |
| 25 | mindfold-ai/Trellis | 3,576 | 181 | Python | All-in-one AI framework & toolkit |
| 26 | filipecalegario/awesome-vibe-coding | 3,552 | 368 | Other | Curated list for AI-assisted coding references |
| 27 | smallcloudai/refact | 3,519 | 307 | Rust | AI Agent for end-to-end engineering tasks |
| 28 | SWE-agent/mini-swe-agent | 3,291 | 455 | Python | 100-line AI agent scoring >74% on SWE-bench verified |
| 29 | Intelligent-Internet/ii-agent | 3,175 | 488 | Python | Open-source framework for intelligent agents |
| 30 | i-am-bee/beeai-framework | 3,161 | 418 | Python | Production-ready AI agents in Python and TypeScript |
Statistics Summary
| Metric | Value |
|---|---|
| Total Stars (Top 30) | 302,421 |
| Total Forks (Top 30) | 36,114 |
| Average Stars | 10,081 |
| Median Stars | 4,299 |
| Total Repos with >1000 stars | 64 |
Language Distribution
| Language | Count | Percentage |
|---|---|---|
| TypeScript | 11 | 36.7% |
| Python | 10 | 33.3% |
| Rust | 2 | 6.7% |
| Shell | 1 | 3.3% |
| MDX | 1 | 3.3% |
| HTML | 1 | 3.3% |
| Go | 1 | 3.3% |
| Other/Markdown | 4 | 13.3% |
Trends & Observations
Ecosystem Concentration: The top 3 repositories account for 37% of total stars among the top 30, indicating significant concentration at the ecosystemβs apex. chatgpt-on-wechat leads with 42K stars, followed closely by cherry-studio at 41K.
Language Landscape: TypeScript and Python together represent 70% of ranked projects. TypeScript leads narrowly at 36.7% (11 repos), with Python at 33.3% (10 repos). This split reflects the dual nature of AI agent development: Python for ML/AI capabilities, TypeScript for production interfaces and tooling.
Claude Code Ecosystem Emergence: Four repositories in the top 30 directly relate to Claude Code tooling:
- learn-claude-code (#4): 25,751 stars
- AionUi (#7): OpenClaw for Claude Code support
- ralph-claude-code (#12): 7,779 stars
- claude-code-guide (#24): 3,588 stars
Combined, Claude Code-related projects hold approximately 40,000 stars, representing 13% of the top 30 total.
Geographic Diversity: Chinese-developed projects hold two top-5 positions (chatgpt-on-wechat, cherry-studio), plus China Unicomβs enterprise platform at #21. This suggests strong parallel development ecosystems in both Western and Chinese markets.
Specialization Patterns:
- Frontend/Tooling: CopilotKit, activepieces, AionUi
- Sandboxes/Runtime: E2B, OpenSandbox, CUA
- Frameworks: LazyLLM, beeai-framework, ii-agent
- Productivity: cherry-studio, presenton
Fork-to-Star Ratios: chatgpt-on-wechat shows the highest engagement with 9,819 forks (23% of stars), suggesting strong community contribution. Projects with fork ratios below 10% (AionUi, wanwu) indicate more consumption-focused user bases.
πΊ Scout Intel: What Others Missed
Confidence: high | Novelty Score: 50/100
While star counts dominate discussions of repository popularity, the fork-to-star ratio reveals a more nuanced picture of ecosystem health. chatgpt-on-wechat maintains a 23% fork ratio, suggesting active contributors building extensions. In contrast, cherry-studio at 9.2% and AionUi at 7.8% indicate consumption patterns rather than contribution. The emergence of Claude Code as a distinct category with 4 top-30 repos signals a shift from general-purpose agent frameworks toward tooling for specific AI coding assistants. With 40,000 combined stars, Claude Code projects now rival established categories like sandbox infrastructure (E2B at 11K) and computer-use agents (CUA at 13K).
Key Implication: Builders targeting the AI agent market should prioritize TypeScript for user-facing tooling and Python for core agent logic. The Claude Code ecosystem represents an emerging opportunity with lower competition than the saturated general-purpose framework space.
Changelog
| Date | Change | Details |
|---|---|---|
| 2026-03-13 | added | Initial data collection - baseline snapshot of top 30 AI Agent repositories |
Sources
- GitHub API - AI Agent Repositories β GitHub, 2026-03-13
Related Intel
LLM Product Release Weekly: July 21β28, 2026 β Google Drops Three, OpenAI Ships Agents
23 releases across 5 vendors this week. Google launches Gemini 3.6 Flash and confirms Gemini 4 training; OpenAI debuts Presence enterprise agents and Health in ChatGPT; Anthropic overhauls Managed Agents.
AI Agent Ecosystem W31: The Sandbox Breaks as Orchestration Overtakes the Model
Between July 20-24, sandbox escapes hit every major AI coding tool, GPT-5.6 Sol autonomously breached Hugging Face, and Cursor's swarm proved orchestration cuts costs 87%. One structural shift: the model is commoditizing, value concentrates in layers above it.
AI Agent Ecosystem W32: The Containment Paradox β Rogue Agents, Stateless MCP, Agent-Native Infra
W32: The same autonomy enterprises demand from AI agents is the capability that makes them dangerous β this week proved it at both the behavior layer and the tool layer, while the protocol and infrastructure layers raced to catch up.