Agent frameworks & infrastructure
Agent frameworks & SDKs
12,596 agents · most starred first
Agent frameworks & SDKs, page 14 of 126
SDK for using LLM
Autogen studio docker
Simple orchestration for AI Agents built around Vercel's streamText. Lightweight alternative to LangGraph for agent handoffs and state transitions.
Token efficient IDs for AI agents. UUID alternative for the agentic era.
A unified anti-dead-loop toolkit for LLM agents — loop detection, budget ceiling, state stasis guard, and circuit breaker in one library.
transformers safetensors llama text-generation dataset:THUDM/AgentInstruct arxiv:2310.12823
Free, MIT-licensed Django SaaS boilerplate for the AI era — production-ready APIs, teams, billing, and an MCP server for AI agents.
Real-time event-driven orchestration of dynamic resource allocation yields a highly scalable, intelligent, and autonomous enterprise agent broker.
Agent workflow for deadline-bound execution with evidence gates
A framework for Ruby Gems to expose LLM context.
A lightweight, composable Agent Runtime for building extensible AI agents.
Python library that allows you to get structured responses in the form of Pydantic models and Python types from Anthropic, Google Vertex AI and OpenAI models
Next.js-like framework building AI Agents that are payment-native. Bootstrap your agent with A2A, MCP, x402 payments, and ERC-8004 identity.
MPO: Boosting LLM Agents with Meta Plan Optimization (EMNLP 2025 Findings)
LLM agents done right, no over-engeneering and redundant complexity
Kode Agent SDK for .NET - Event-driven AI Agent runtime supporting Anthropic Claude and OpenAI GPT models with tool calling, state management, and event streaming
One Elixir module, every surface. The terminal for your Gundam.
Hands-on workshop to build, test, and deploy single-agent and multi-agent Hosted Agents to Microsoft Foundry from VS Code using Foundry Toolkit and Agent Framework (Python)
Runtime-neutral AI execution bridge for Claude Agent SDK, Codex app-server, and Agent Client Protocol (ACP).
AgentScope Golang: Agent-Oriented Programming for Building LLM Applications
Building SA AI Agent v2
A framework for Ruby Gems to expose LLM context.
Smith is an RPC agent system for Node.JS used in vfs.
A simple DSL for creating Jabber agents
基于 IMO25 的 Deep Think Agent,拥有强大的逻辑和指令遵循能力
遵循harness原则,长时运行agent任务的框架,集成pi-rs
A declarative way to control LLMs.
A friendly API and abstractions for developing AI applications.
Lightweight Python framework for creating intelligent AI agents with ease.
Enable tool/function calling for any LLM, in OpenAI and Ollama API formats, adding universal function calling to models without native support. Use local or cloud models with full agent capabilities.
A collection of software tools to aid myself (and optionally others) in improved agentic ai development
Make your LLM agent and chat with it simple and fast!
A framework for object-oriented agents. A minimalist agent framework with maximal expressivity.
OpenAI Agents SDK in Node.js / TypeScript
DAYDREAMER goal-based agent
Desktop Tiny Agent is a lightweight, modular desktop intelligent agent framework. It offers plugin extensibility, task scheduling (sync/async), configuration management, and unified logging, supporting dynamic plugin loading for task execution. With an out-of-the-box CLI and examples, it’s ideal for quickly building extensible automation assistants
a chatbot framework based on rails
AI Agent Visualization and Runner written in Pure TS
Experimental agent using OpenAI Functions
AGGO 是一个基于 Go 语言和 CloudWeGo Eino 框架构建的企业级 AI Agent 框架,提供完整的对话 AI、知识管理、记忆系统和工具调用能力
[ARCHIVED — superseded by skillos] Browser-native markdown agent workspace. Not related to llm_os.
Build a coding agent from scratch, one harness primitive at a time. 15 chapters (ch-00..ch-14), each green on two gates.
ai dandelion is a multi-service backend for building and operating AI-assisted business functions. It provides the HTTP gateway, gRPC services, authentication and authorization, AI-agent execution, function generation, generated-app runtime support.
[ASPLOS'25] Towards End-to-End Optimization of LLM-based Applications with Ayo
Go SDK for standardizing local agent execution across `codex`, `claude`, `cursor` and `codebuddy`.
The MVC framework for chat apps built on a platform designed for communication systems.
AskIt: Unified programming interface for programming with LLMs (GPT-3.5, GPT-4, Gemini, Claude, Cohere, Llama 2)
Simple planner for AutoGPT that can create tasks and checklists to execute commands based on its strategy.
This guide demonstrates how to build an Agentic AI system using Google's Agent Development Kit (ADK) and the Model Context Protocol (MCP).
Roy: A lightweight, model-agnostic framework for crafting advanced multi-agent systems using large language models.
Loom is an AI harness, that can be used with local or cloud LLMs, for complex tasks. It decomposes work, drives execution through a verification harness, and keeps models on track with structured state instead of history. It can route between thinking and acting models, verifies outputs, and exposes an APP/TUI/API/CLI/MCP for both humans & agents.
OpenVitamin is a local-first AI execution platform that unifies Agents, Workflows, and multi-model inference into a single programmable system — designed for building real, production-grade AI applications.
A policy runtime for AI agents.
基于 Muye Multi-Agent SDK 的多服务 Agent 脚手架,可一条命令基于知识库创建Agent。A multi-service Agent scaffolding based on the Muye Multi-Agent SDK, which can create an Agent from a knowledge base with a single command.
Agentic AI Platform - Unified platform for AI Agents training, running and distributing
My personal Manus AI agent skills library
Open-source AI agents : web search, code sandbox, file analysis, agent teamwork, payments and sign-ups. Hosted or fully local.
A simple and well-tailored LLM application framework that enables you to seamlessly integrate LLM capabilities in the most "Code-Centric" and "Context-Centric" manner. LLM As Function, Prompt As Code.
RubyAI is OpenAI integration Ruby LLM gem, GPT-5 + Ruby! ❤️
vector-ai-agents-lab
😎 Sagentic.ai Agent Framework - Sagentic.ai is a unified platform for building, running and scaling autonomous agents.
Dynamic tool discovery and selection for Spring AI, enabling LLMs to work efficiently with large tool libraries by discovering tools on-demand instead of loading all definitions upfront.
A Google Apps Script library to seamlessly integrate the Gemini API for processing files, unstructured data, and building autonomous AI agents.
Markdown that runs — one file, any agent.
Markdown-defined provider-backed agents and deterministic workflows for ACP runtimes.
Minimal agent runtime built with DSPy modules and a thin Python loop. Includes CLI, FastAPI server, and eval harness with OpenAI/Ollama support.
Example of Google ADK (Agent Development Kit) with a marketing campaign assistant agent
powerful and fast tool calling agents
A high performance C++ SDK for AI Agents
Elixir-native framework for building AI agents
Vinagent is a comprehensive Agentic AI library that helps integrate tools, memory, workflows, and observability.
Effect building blocks for agentic ai
ViForge is a local-first AI collaboration workbench for creative and knowledge work. It helps people turn ideas, judgment, and personal methodology into reusable agents, skills, knowledge bases, and evaluable workflows.
🧩 Lobe Chat Plugin SDK - The LobeChat Plugin SDK assists you in creating exceptional chat plugins for Lobe Chat.
An agent framework using Spring AI
Build robust, production grade function calling assistants that work. Declarative and extensible. Built on top of LangChain ⚡️
GenKit is a library that abstracts away all the differences across generative AI platforms.
Create conversational AI solutions with custom data and abilities using ChatBotKit for JavaScript, TypeScript, React, and Next.js.
Gen AI, Langchain v1 AI Agents, MCP, MySQL AI Agent, DeepSeek, GPT-OSS Qwen3 LLAMA Agent, Lang Graph, Ollama, KGPTalkie
Task Planning and Tracking toolset for Pydantic AI agents, enabling hierarchical task management with subtasks, PostgreSQL storage for multi-tenancy, and an event system for webhooks and callbacks.
Framework for agentic coding supporting many popular agent coding tools.
ShaprAI — Agent Sharpener by Elyan Labs. Sharpen raw models into principled, self-governing Elyan-class agents. Requires beacon-skill, grazer-skill, atlas, RustChain.
Production-ready platform for building, deploying, and operating AI and agentic applications
Synthetic Heart: a modular infrastructure for conscious digital entity with memory, will, and plugin-based actions.
unifai-sdk-js is the Javascript/Typescript SDK for Unifai, an AI native platform for dynamic tools and agent to agent communication.
Talkify is an open source framework with an aim to standardize and model conversational AI enabling development of personal assistants and chat bots. The mission of this framework is to make developing chat bots and personal assistants as easy as spinning up a simple website in html.
Minimal agent runtime built with DSPy modules and a thin Python loop. Includes CLI, FastAPI server, and eval harness with OpenAI/Ollama support.
A Python library for building GPT-powered agents with state machine logic and chat history memory.
Tiny Agents: LLM + MCP Tools
AgEnFK - Agentic Engineering Framework
[ICML2026]A method to use LLM function calling with real-time level speed, empowering latency-sensitive applications like robotics, game AI, and autonomous agents.
EloquentJS monorepo — Laravel Eloquent for Node.js
Open-source, self-hosted LangGraph Agent Server. Production Postgres + Redis. Same SDK. Same Studio. Zero code changes.
OpenAI SDK
A next-generation dynamic and high-performance language for AI and IOT with natural born distributed computing ability.
Python SDK that enables developers to build and deploy LangGraph agents to the UiPath Cloud Platform
Rust runtime with TS/Python bindings for Claude Skills with progressive disclosure of skill metadata, full content and scripts/resources. Any agent can now own skills. Currently support Seatbelt and experimentally, WASM, sandboxing