Trending AI projects list

A curated overview of the top open-source AI repositories, organised by category. Each section covers a distinct layer of the modern AI stack — from orchestration frameworks and RAG pipelines to coding agents, evaluation tooling, and model serving infrastructure. This list is subject to change and will be updated regularly.

Maintained at bobpage0451/ai-trending-repos • Last updated: 2026-07-13 (refreshed bi-weekly)

• App orchestration layer
• RAG, memory, and knowledge layer
• Language models and foundation model layer
• Agent and workflow layer
• Tools, MCP, and integration layer
• UI and generative interface layer
• AI coding agents and developer experience
• Evaluation and testing layer
• Observability, tracing, and prompt management
• Guardrails, safety, governance, and policy layer
• Structured output and prompt engineering layer
• Document ingestion, parsing, and data preparation
• Model training, fine-tuning, and adaptation
• Deployment, local inference, and model serving layer

---

App orchestration layer

This is the core application framework layer for building AI apps. It helps structure LLM calls, chains, workflows, tool use, retrieval, memory, callbacks, and agent execution.

Connect prompts, models, tools, memory, retrieval, agents, and workflows.

Repository	⭐ Stars	🍴 Forks	License	Language
langchain-ai/langchain	141.2k	23474
appsmithorg/appsmith	40.3k	4639
langchain-ai/langgraph	36.7k	6161
langchain-ai/deepagents	25.9k	3632
deepset-ai/haystack	25.8k	2901		MDX
google/adk-python	20.5k	3661
Chainlit/chainlit	12.3k	1726

RAG, memory, and knowledge layer

This layer includes retrieval-augmented generation, semantic search, embeddings, vector storage, hybrid search, reranking, citations, document permissions, and short-term or long-term memory.

Give the model external, project, user, or company knowledge.

Repository	⭐ Stars	🍴 Forks	License	Language
infiniflow/ragflow	84.5k	9851
mem0ai/mem0	60.3k	7001
MemPalace/mempalace	57.1k	7373
run-llama/llama_index	50.7k	7707
milvus-io/milvus	45.1k	4108
VectifyAI/PageIndex	33.9k	2962
qdrant/qdrant	33k	2464

Language models and foundation model layer

This layer includes proprietary, open-weight, and local language models used for chat, reasoning, coding, embeddings, multimodal tasks, and agentic workflows. It covers model families such as OpenAI GPT, Anthropic Claude, Google Gemini and Gemma, Meta Llama, Mistral, DeepSeek, Qwen, Cohere, GitHub Models, and other foundation models.

Choose, compare, and use the underlying language models that power AI applications.

Repository	⭐ Stars	🍴 Forks	License	Language
deepseek-ai/DeepSeek-V3	103.9k	16729
google-deepmind/gemma	5.5k	986

Agent and workflow layer

This layer covers tool-using agents, graph-based workflows, planner-executor patterns, multi-agent collaboration, human approval steps, state management, retries, and long-running task execution.

Build systems that can plan, call tools, perform multi-step tasks, and coordinate agents.

Repository	⭐ Stars	🍴 Forks	License	Language
openclaw/openclaw	382.1k	80165
langflow-ai/langflow	151.3k	9469
bytedance/deer-flow	76.4k	10357
FoundationAgents/MetaGPT	69.2k	8830
crewAIInc/crewAI	55.1k	7745
agno-agi/agno	41k	5605
OpenBMB/ChatDev	33.7k	4196

Tools, MCP, and integration layer

This layer includes Model Context Protocol servers and clients, tool schemas, function calling, API connectors, filesystem access, browser access, database access, and integrations with services like GitHub, Slack, Google Drive, and internal tools.

Let models and agents interact with external systems, APIs, databases, files, and tools.

Repository	⭐ Stars	🍴 Forks	License	Language
google-gemini/gemini-cli	105.8k	14226
browser-use/browser-use	103.3k	11436
D4Vinci/Scrapling	68.5k	6786
microsoft/playwright-mcp	34.8k	2885	![Apache-2.0](https://img.shields.io/badge/license-Apache%20