Agent frameworks & infrastructure
Agent frameworks & SDKs
12,596 agents · most starred first
Agent frameworks & SDKs, page 41 of 126
A devel of production-grade Agentic AI systems including autonomous research agents, coding agents, RAG pipelines, and job search agents built with Python, LangChain, LangGraph, FAISS, and LLMs for real-world deployment.
A modern open-source starter template for building AI-powered Todo agents with Next.js, LangGraph, LangChain, MongoDB, and TypeScript. Learn clean architecture, tool calling, memory, streaming, and production-ready AI patterns.
一个轻量化的ai agent框架
Formal framework for governing autonomous AI agents through resource constraints, budgets, and lifecycle management
Enterprise-Grade Fault Tolerance and Resilience for AI Agents - circuit breakers, bulkheads, retries, and model fallbacks.
Five hero examples of effectful AI agents built on Effect v4 (beta). Pure Effect programs; one local entrypoint; one Cloudflare Worker hosts all five.
Fault tolerant, event driven, extensible AI Agent framework
Substructure is an open-source engine that drives your agent loop with JSON webhooks. It keeps sessions durable and streams AG-UI events to your clients. It's language agnostic. You don't even need an SDK.
Agent as a Service (AaaS) - Enterprise AI Agent Platform | Deploy AI Agents at Scale | Contact: @wilbtc on Telegram
A monorepo boilerplate for building AI chatbot agents with TypeScript, NestJS, and React
Magic tool to generate things
reShapr Community repo
Production-grade AI agents, RAG systems, and autonomous model-agnostic workflows using .NET 10, Microsoft.Extensions.AI, and Semantic Kernel. Features dynamic dual-engine routing (Gemini & Groq/Llama), structured JSON enforcement, and an Angular 19 cyber-dashboard UI.
Examples demonstrating how to build, route, and manage stateful AI agent graphs from scratch using LangGraph.
A programming language designed for ai agents — minimal tokens, canonical syntax, batteries included.
A collection of sample agents built with ADK-TS (Agent Development Kit for Typescript)
有情感的自主进化神经感知多agent代理框架
Ai-Agent Builder is a production-grade, "backend-less" single-file web application for authoring, configuring, and packaging AI agents. It runs entirely in the browser — no server, no API keys stored, no telemetry — making it invisible to trackers and safe for enterprise environments.
Agent orchestration
Give your Python objects the ability to think and act — no glue code, no prompt engineering.
An open-source AI agent platform that does the work for you.
Thin Integration Harness Engine for SMEs — decouple Shell, Brain, and plugins via OpenHarness. Headless, contract-first orchestration. Not an Agent framework.
QuickForge AI: A production-ready fullstack template combining FastAPI (Python) and TypeScript for rapid AI prototype development. Designed for freelancers and AI entrepreneurs who need to quickly build and deploy professional applications with modern development practices, testing frameworks, and CI/CD pipelines.
An open framework for cognitive programming with LLMs: structured, reusable, inspectable workflows built from Data Units, Operations, Modules, Contracts and Traces.
Agent Run Config — an open specification for declaring, packaging, securing, and sharing portable, governed AI agents. Like a Dockerfile for agents: one reviewable Agentfile for identity, tools, boundaries, policy, and OCI packaging.
Open-source Go agent runtime for autonomous AI systems and multi-agent automation. MCP, RAG, 15+ LLM providers and 60+ built-in tools in a single binary.
A modular AI Agent framework designed for complex task orchestration and logic reasoning.
M.A.R.I.A. — local autonomous cognitive architecture, offline-first, modular, safety-aware.
🚀 Transpile Ion to C efficiently, leveraging move-only ownership and channels for safe concurrency in a modern systems programming environment.
Productized LLM Wiki pattern + typed pack registry + proposal-only governance — UNESCO Memory of the World Triple Anchor
Proprietary, all-rights-reserved operating layer for governed AI applications — policy, authority, evidence, execution, and recovery. No general license is granted; see LICENSE.
FastAPI server and CLI for Agentflow. Auto-generates a REST API from your agent graph, with JWT/custom auth, RBAC authorization, rate limiting, dual-layer checkpointing, and Docker/Kubernetes-ready builds.
Chat with AI agents to build, manage, and orchestrate custom tools, dashboards, and it's all powered by cats.
A lightweight agent framework. Start with the core, add only what is needed.
Framework-agnostic Go package for integrating AI-powered chat functionality into any web application. It features out-of-the-box support for popular Go web frameworks, a flexible model abstraction for using OpenAI, Anthropic, xAI, Google Gemini, Meta, and more.
Runnable LangGraph workflow examples covering conditional routing, parallel execution, iterative refinement, approvals, and checkpoint recovery.
Simple LLM agent written in Go
Native Swift package for multi-provider LLM chat, streaming, tool calling, and vision.
GCF Python implementation. 100% LLM comprehension on every frontier model. 50-92% fewer tokens than JSON. 43B+ round-trips verified. Zero dependencies.
Native agent-graph runtime written in C++20
Graph-based runtime for building, running, and debugging LLM agents in Go.
Open Symbolic AI Core Repository
ai4X turns human intent into fit-for-purpose Agentic Constellations—coherently curated and explicitly activated under human authority.
Native C++17 AI agent runtime — local-first, embeddable, MCP-native. Tool-calling, memory, and an async agent loop that runs where Python can't: ROS2, game threads, drones, HFT.
docker region:us
An open-source framework for building stateful, multi-agent AI applications. It helps developers create reliable AI workflows with memory, human-in-the-loop capabilities, tool calling, and long-running agent orchestration.
transformers safetensors qwen3 text-generation agents tool-use
Reference Runtime for Intent-Oriented Programming (IOP). Data declares what, runtime decides how.
Agent-native action layer for Laravel with shared human and agent surfaces, approvals, and audit logging.
Build and run ArchAgents solutions from your coding agent, with deployable examples and developer tooling.
Lightweight, decorator-driven Express + TypeScript backend starter — clone, run, ship. AI-agent ready.
Pattern for emulating named custom subagents with the Amp SDK: dynamically defined roles with their own system prompts, scoped tool permissions, and isolated context per invocation.
GCF Rust implementation. 100% LLM comprehension on every frontier model. 50-92% fewer tokens than JSON. 43B+ round-trips verified. Zero dependencies.
Single-user, local agentic harness
Loop Engineering framework: design autonomous agent feedback cycles instead of hand-prompting each step
Cognitive OS v3.0 — Universal AI agent operating system specification. 16-section cognitive architecture for Claude Code, Cursor, Codex, Gemini, and Factory Droid. Multi-harness prompt engineering with verification chain, swarm coordination, and adversarial security.
Zero-error LLM execution via SPRT voting. Rust library and MCP server implementing the MAKER algorithm for mathematically-grounded error correction in long-horizon AI agent tasks. Research experiment based on arXiv:2511.09030
Run an AI company. Not a prompt. Open-source, privacy-first desktop app for AI-agent organizations — Windows, macOS, Linux.
Convos API: auth, credits, subscriptions, agent templates, and notifications.
Agent reasoning loop with tool calling for Gemma 4 models. Handles prompt construction, tool call parsing, execution, and multi-turn conversation management.
Define LangGraph agent workflows in JSON. DLG handles validation, compilation, resilience, and execution — no hardcoded graph topologies required.
Cognitive runtime layer for AI engineering. Adds planning, memory hierarchy, and deterministic execution on top of any language model.
A flask extension to build and deploy chatbots to multiple channels
These are the Microsoft Semantic Kernel Workshop notebooks. It addresses AI agents, agents collaboration and, of course, kernel, plugins, planners and function calling
A (nearly) fully featured and up-to-date c++ clone of the openai-node library.
Build enterprise AI agents with DAG scheduling, hierarchical memory, 20+ tools, and multi-level permissions for safe local or sandboxed execution
You own this harness as a stack of installable workers, and “build your own” stops meaning “fork a framework.” It means “swap a few workers.”
One-file Python SDK for OpenAI-compatible LLM APIs: streaming + reasoning, tools from callables, strict structured output, vision/audio/video/files.
One-click Render deploy template for nanobot — a lightweight self-hostable AI agent
Build self-evolving AI agent harnesses with portable harness units, artifact-aware testing, trace-backed diagnosis, and evidence-gated promotion.
Independent LangGraph C++ runtime for stateful AI agents and recoverable graph workflows: checkpointing, streaming, interrupts, tool calling, edge/local LLM apps. C++23 developer preview; not official LangGraph/LangChain.
Sovereign AI Initiative for Open Autonomous Cognitive Companion
Starter Kit for AI Native Applications - Vite Plus (Vite+) TanStack Start monorepo with Paraglide.js (i18n), drizzle-orm, better-auth, Feature-Sliced Design (FSD)
OCaml-first LLM SDK with a zero-dependency npm package: OpenAI, Anthropic, xAI, Meta, OpenRouter and local models; typed chat, tools, structured outputs, streaming, cost tracking.
Rust based JSON repair library
MUNDO - THE EMPEROR. Complete AI orchestration system with 1208 skills, 25 capability modules, self-evolving, collective consciousness. GitHub Actions 24/7 automation.
Self-hostable AI agents you can actually leave running — durable crash-resume, governed autonomy, seccomp sandbox, offline-replayable workflows, CI-native eval. Async-Python, YAML-config, multi-provider (OpenAI/Anthropic/Cloudflare), MIT.
TypeScript SDK for writing AI agent plugins once and exporting them to multiple agent platforms.
Agnostic AI Agentic System
[In progress] The useless AI agent.
Bun-native CLI agent: long-term memory, knowledge graph, multi-channel (Telegram/Slack/Discord), 25+ LLM providers.
Build an agent in minutes. Add production controls when you need them.
Agent Lifecycle Framework: Documentation as executable infrastructure for birthing and evolving autonomous AI agents
Error handling and recovery patterns for improved reliability of Pydantic AI Agents.
An Agent-First Repository is a human-governed software environment designed so AI agents can execute engineering work efficiently while deterministic controls constrain their authority, verify their output, and preserve human accountability.
Experimentation system for autopoietic, self-modifying agent organisms — a decentralized agent loop with persistent corpus, endogenous self-extension, and emergent governance. Anyone who views AGI not merely as a “very clever machine” but as genuine cognition can hardly ignore autopoiesis.
minorAgent —— 轻量、清晰的 Agent 框架。内置丰富工具与多 Agent 协作,强可扩展性,适合 Agent 入门学习与二次开发。期待与社区共建。
Questionnaire-style batch questioning plugin for DeepSeek Harness: 10+ questions at once (single/multi/yes-no toggle/compare/open), per-question skip, fullscreen overlay, two-column recap after submit
🧾 Agentspecs — YAML-based specifications for AI agents, MCP servers, Skills, Guardrails and more...
多租户 SaaS 智能体
Nulang is an actor-based runtime and programming language that's type-safe, compiled, and designed for building fast, distributed, resilient applications.
Agent harness / control-plane around a stateless LLM: deterministic policy floor on PreToolUse/PostToolUse hooks, secret redaction, injection fencing, tool-result compression, code-graph memory, orchestration. Five plugin modules for Claude Code (Tier 1) or any MCP host (Tier 2). MIT.
Universal harness-free AetherMind continuity primitive for agent work.
IDE for authoring, deploying, and monitoring BDI-based AI agents that run as automated cron jobs. Transform agent definitions (AGENT.md with beliefs/schema, desires/goals, intentions/constraints) into persistent, self-executing AI agents.
Self-improving AI agent harness in Zig 0.16 — gated self-patch loop, sandboxed WASM tools, CLI + web UI
Production-ready full-stack template for building AI agents with LangGraph, LangChain & DeepAgents — FastAPI + Aegra backend with JWT auth and a Next.js 16 frontend.
Manusclaw: Unleash self-reasoning CLI beasts to execute code, browse the web, and dominate tasks — across 12+ messaging channels, with voice wake, live canvas, SSH control, and multi-agent routing. No limits. Pure power.
An Effect-native agent execution kernel: sessions, runs, turns, steering, follow-ups and typed lifecycle events on top of Effect AI.
The full-stack TypeScript framework for AI agents. An agent is a file: drop it in agents/ and it gets an HTTP route, token streaming, human-in-the-loop approvals and a typed client. React pages with file routing and SSR, OAuth sessions, Postgres, WebSockets, cron, and nine deploy targets. Apache-2.0.