AI Tools & Repos

TypeScript toolkit for building AI-powered products. Unified API for OpenAI, Anthropic, Google, and more with streaming support.

Microsoft's SDK for integrating LLMs into apps. Supports C#, Python, and Java with enterprise-grade features.

Programming framework for foundation models. Replace hand-written prompts with composable, optimizable modules.

End-to-end LLM framework by deepset. Build production-ready RAG, question answering, and semantic search pipelines.

Framework for orchestrating role-playing AI agents. Define agents with specific roles, goals, and tools to work as a team.

Lightweight multi-agent orchestration framework from OpenAI. Focuses on handoffs and routines between agents.

Agent framework by the creators of Pydantic. Type-safe, model-agnostic agents with structured output validation.

Library for building stateful, multi-actor LLM applications as graphs. Supports cycles, branching, and persistence.

The original autonomous AI agent. Give it a goal and it plans, executes, and iterates with minimal human intervention.

Multi-agent framework that assigns different roles (PM, architect, engineer) to collaborate on complex software tasks.

Official collection of MCP servers for popular services — GitHub, Slack, Google Drive, PostgreSQL, filesystem, and more.

Anthropic's agentic coding tool. Uses MCP for extensibility — connect to databases, APIs, and custom tools via MCP servers.

AI-powered code editor with MCP support. Integrates Claude, GPT-4, and custom models for code completion, chat, and editing.

Registry and marketplace for MCP servers. Discover, install, and manage community-built MCP servers with one command.

Open-source AI code assistant for VS Code and JetBrains. Supports any LLM with autocomplete, chat, and edit features.

GitHub's AI pair programmer. Code completions, chat, and now agent mode for multi-file edits across your workspace.

Autonomous coding agent for VS Code. Plans, writes, and debugs code with human-in-the-loop approval for each step.

AI-native code editor by Codeium. Features Cascade for multi-file agentic editing and Supercomplete for intelligent completions.

AI full-stack web app builder by StackBlitz. Describe your app in natural language and get a complete, deployable project.

AI-powered UI generation tool. Describe a component or page and get production-ready React/Next.js code with Tailwind CSS.

The first AI software engineer by Cognition. Handles complete engineering tasks: planning, coding, testing, and debugging.

Fine-tune LLMs 2-5x faster with 80% less memory. Supports Llama, Mistral, Gemma, and other popular architectures.

Streamlined tool for fine-tuning LLMs. YAML-based configuration for LoRA, QLoRA, full fine-tuning, and DPO/RLHF.

HuggingFace library for parameter-efficient fine-tuning. LoRA, QLoRA, prefix tuning, and more — fine-tune with minimal resources.

Transformer Reinforcement Learning library. SFT, reward modeling, PPO, DPO, and ORPO training for LLMs with HuggingFace.

Evaluation framework for RAG pipelines. Measures faithfulness, answer relevancy, context precision, and context recall.

Test and evaluate LLM outputs. Compare prompts, models, and RAG pipelines with automated testing and CI integration.

Serverless vector database built on Lance format. Embedded, zero-config, and supports multi-modal data with full-text search.

Open-source AI-native embedding database. Simple API for storing and querying embeddings with metadata filtering.

Open-source vector database with built-in vectorization. Hybrid search (vector + keyword), GraphQL API, and multi-tenancy.

LangChain's platform for debugging, testing, evaluating, and monitoring LLM applications. Deep integration with LangChain/LangGraph.

Open-source AI observability by Arize. Trace LLM apps, evaluate RAG quality, detect hallucinations, and monitor in production.

High-throughput LLM inference engine. PagedAttention for efficient memory management, up to 24x higher throughput than HuggingFace.

Unified API proxy for 100+ LLM providers. Call OpenAI, Anthropic, Bedrock, Vertex, and local models with one interface.

Unified API for accessing many LLMs through a single endpoint. Automatic fallback, usage tracking, and competitive pricing.

Desktop app to discover, download, and run local LLMs. Beautiful GUI with chat interface and OpenAI-compatible server.

Open-source ChatGPT alternative that runs 100% offline. Download and run models locally with a clean, modern interface.

Ultra-fast LLM inference powered by custom LPU hardware. Run Llama, Mixtral, and Gemma at industry-leading speed.