A strictly vetted directory of open-source artificial intelligence repositories, local LLM engines, multi-agent frameworks, and vector search tooling. Annotated with personal engineering takes from production builds.
Curated Open-Source AI Repositories
All Categories
MCP & Tooling
LLM SDKs
Inference & Serving
Agent Frameworks
RAG & Vector Search
Video & Media
Automation
Search & Grounding
LLM Tooling
Content
Workflow Orchestration
Data & Storage
Learning
AI Operations
Infrastructure
MCP & TOOLING
modelcontextprotocol/servers ↗
★ 18,500
Reference implementations for Model Context Protocol (MCP) servers (Postgres, SQLite, Git, Filesystem, Slack).
Why It's Essential:
“The definitive gold standard for how LLM agents interact with local databases, system tools, and APIs cleanly without messy custom glue code.”
AGENT FRAMEWORKS
openclaw/openclaw ↗
★ 302,000
Personal AI assistant for cross-platform task execution — 302k stars, the 2026 breakout for agentic execution.
Why It's Essential:
“Run on your hardware in India, keep data onshore for DPDP — personal assistant that ships tasks without API lock-in. Top trending 2026 per NocoBase/ByteByteGo.”
LLM SDKS
anthropics/anthropic-sdk-python ↗
★ 4,200
Official Python client library for Anthropic Claude models, tools, and batch processing.
Why It's Essential:
“Remarkably clean async client with native prompt caching and tool-use support that outperforms raw HTTP requests in production.”
AGENT FRAMEWORKS
langchain-ai/langchain ↗
★ 132,476
Foundational agent framework — chains, tools, agents — 132k stars, still the center of tutorials in 2026.
Why It's Essential:
“Base for Laravel AI SDK, RAGFlow, LangGraph — essential India stack behind Dify/Langflow. Swiss-army for Gujarat SME RAG.”
LLM SDKS
google-gemini/generative-ai-python ↗
★ 5,800
Official Python SDK for the Google Gemini API, structured outputs, and multimodality.
Why It's Essential:
“Massive context windows (up to 2M tokens) and native image/audio understanding make this essential for deep analysis agents.”
WORKFLOW ORCHESTRATION
langgenius/dify ↗
★ 136,278
Production AI app platform — workflows, RAG, agents — 136k stars, enterprise-ready with visual builder.
Why It's Essential:
“Gujarat SMEs ship WhatsApp + RAG flows without code — Dify + n8n + pgvector is the ₹500 VPC stack.”
INFERENCE & SERVING
vllm-project/vllm ↗
★ 36,000
A high-throughput and memory-efficient inference and serving engine for LLMs using PagedAttention.
Why It's Essential:
“The backbone of self-hosted open LLM deployments. 2-4x higher throughput than HuggingFace TGI.”
WORKFLOW ORCHESTRATION
langflow-ai/langflow ↗
★ 146,595
Visual drag-and-drop builder for agents and RAG — 146k stars, built on LangChain.
Why It's Essential:
“Prototype Hindi/Gujarati RAG graphs visually — fastest from idea to demo for Junagadh teams.”
AGENT FRAMEWORKS
langchain-ai/langgraph ↗
★ 12,500
Build resilient language agents as graphs with cyclical loops, human-in-the-loop, and persistence.
Why It's Essential:
“Provides state-machine determinism to AI agent loops instead of leaving multi-step workflows to random chance.”
RAG & VECTOR SEARCH
run-llama/llama_index ↗
★ 39,000
Context augmentation framework for LLM applications with advanced RAG indexing and retrieval.
Why It's Essential:
“Best-in-class data connectors for PDFs, Notion, SQL, and enterprise data with hybrid retrieval.”
VIDEO & MEDIA
remotion-dev/remotion ↗
★ 21,000
Make programmatic videos with React and CSS motion design.
Why It's Essential:
“Programmatic video rendering engine that turns React components into frame-accurate 60fps MP4 reels.”
DATA & STORAGE
supabase/supabase ↗
★ 98,900
Postgres-based data platform for web/mobile/AI — 98.9k stars, open-source Firebase.
Why It's Essential:
“Gujarat VPC Postgres + pgvector + Auth + Storage in one — run RAG locally at ₹0 token cost, keep DPDP data onshore.”
AUTOMATION
n8n-io/n8n ↗
★ 56,000
Fair-code workflow automation platform with native AI agent nodes and LangChain integrations.
Why It's Essential:
“Visual workflow canvas with full self-hosting control and native AI agent routing.”
INFERENCE & SERVING
ollama/ollama ↗
★ 115,000
Get up and running with Llama 3, Mistral, and other large language models locally.
Why It's Essential:
“The Docker for local LLMs. Run 8B and 70B parameter open-weights models on Apple Silicon with 1 command.”
LLM TOOLING
BerriAI/litellm ↗
★ 19,500
Call 100+ LLMs using the OpenAI Input/Output Format with proxy routing, fallbacks, and cost tracking.
Why It's Essential:
“Unified interface for OpenAI, Anthropic, Gemini, Groq, and Bedrock with automatic fallback switching when an API is down.”
AGENT FRAMEWORKS
browser-use/browser-use ↗
★ 31,000
Make websites accessible for AI agents via Playwright and visual vision models.
Why It's Essential:
“Enables agents to interact with modern complex SPAs, solve captchas, and execute browser workflows autonomously.”