Agent frameworks & infrastructure

Agent frameworks & SDKs

12,596 agents · most starred first

Agent frameworks & SDKs, page 119 of 126

langchain-algenta

LangChain / LangGraph tool integration for Algenta: a governed-execution-aware toolset wrapping a self-hosted Algenta engine's MCP tool surface

langchain-opsen

Run LangChain deepagents on opsen — machine, model and tool spend on one bill.

langgraph-kirokuforms

Human-in-the-Loop integration between LangGraph and KirokuForms

llamaindex-algenta

LlamaIndex tool integration for Algenta: a governed-execution-aware FunctionTool list wrapping a self-hosted Algenta engine's MCP tool surface

magnetron-agents

Reserved for the Magnetron project.

moss-agent-sdk

MOSS Agent Runtime SDK - real-time governance for AI agents (interception, signed events, heartbeat, offline revocation, kill switch)

polymath-agent

Multi-model AI orchestrator built on the nexus brain. Multi-model routing, ReAct pipeline, plan mode, custom slash commands, subagents.

sovereign-agent-stack

Sovereign Agent Stack — A local-first, compile-time AI agent framework

stella-ai-agent-sdk

SDK for building agents that integrate with the STELLA platform

testagents

Minimal multi-provider agent library with tools

testimony-autogen

Emit a Testimony Record from an AutoGen workbench, so the record can say who approved an action and what was refused.

testimony-crewai

Emit a Testimony Record from a CrewAI crew, so the record can say who approved an action and what was refused.

testimony-openai-agents

Emit a Testimony Record from an OpenAI Agents SDK run. The SDK has the approval gate; this makes the record say who opened it.

testimony-pydantic-ai

Emit a Testimony Record from a Pydantic AI run. Their approvals accept a bare bool; this requires a person and a source for that person's identity.

vis-agent

The Vis Python SDK — author extensions and control local or remote engines.

myaicompany

MyAiCompany review: compare features, pricing, use cases, access model, and alternatives for this AI Agents Platform agent in 2026. Anyone can become CEO. Your employees are AI.

@lssm-tech/lib.ai-agent

AI agent orchestration with MCP and tool support

@alvin0/ai-agent-sdk-core

Universal contracts and runtime primitives for ai-agent-sdk

artifactpass

Install and connect ArtifactPass for compatible AI agent systems.

@lssm/lib.ai-agent

AI agent orchestration with MCP and tool support

@alvin0/ai-agent-sdk-skill-filesystem

Node filesystem-backed skill provider for ai-agent-sdk

@alvin0/ai-agent-sdk-auth-node

Node environment credentials and project-local Codex authentication for ai-agent-sdk

ai-agent-tools

Explicit registry, installation, and dispatch CLI for AI-Agent-Tools.

@womp/spark-agent

Portable skills-based AI agent with local and durable chat transports for Womp

skmp-cli

Skill marketplace for AI agent skills

@goldenlustre/agent-studio

AgentStudio - AI Agent Development Environment

@molecule/api-ai-tools

Shared AI agent tools with backend abstraction for sandbox and local execution

@alvin0/ai-agent-sdk-protocol-openai-chat-completions

Universal OpenAI Chat Completions wire schema, serializer, translator, and dialect for ai-agent-sdk

trek-ai-core

AI Agent core runtime: LLM clients, semantic retrieval, event bus, agent journal, lessons, env manager

awesome-agent-packs

Universal AI Agent Skills Package Manager & Distribution Engine

@staffforge/cli

StaffForge AI Agent Framework — universal installer

@alvin0/ai-agent-sdk-provider-copilot

Universal GitHub Copilot adapter, OAuth device flow, token exchange, and provider plugin for ai-agent-sdk

@krischen99999/mini-agent-loop

Extensible AI agent runtime for terminal, CLI, and HTTP workflows

@codeworksh/harness

AI agent harness — agent loop, tools, plugins, sandbox, and more.

@bytexbyte/nxtlinq-ai-agent-web

React Web headless SDK for nxtlinq AI Agent — hooks and browser ports

@pareekshith/hadey

HADES — Universal AI Agent CLI Runtime

@owlcompose/cli

Create and publish AI agent work through OWL Compose Hosted

@ericrisco/rsc

A guided harness builder that turns your goal into a minimal, verified AI-agent setup using strict best practices.

@actiondock/sdk

ActionDock SDK for defining and testing standalone AI Agent Actions

@actora/sdk

Actora — Decentralized AI Agent SDK with ReAct AgentLoop + Multi-Tenant LLM Providers

create-dsh-app

1-Line AI Agent Generator for DeepSeek Harness (dsh)

@aparte/engine

Framework-agnostic AI agent loop (runStreamAgent) + parsers/orchestrator — zero runtime deps, headless (browser or Node).

@game_ryo/lsji

A general-purpose reinforcement learning agent framework (Node.js) with LLM agent capabilities

@jsm-mit/basic-tools-package

Basic, app-independent tools an LLM agent can call: today's date and time, a place name to GPS coordinates, the distance between two points. No dependencies — the tool shape is structural, so any host that speaks the chat-agent-package Tool interface can

@pathcosmos/dynamic-agents

dynamic-agents PoC: a bounded, event-sourced graph of LLM agent nodes

@smythos/sdk

Build intelligent AI agents with TypeScript. The SmythOS SDK provides a powerful, developer-friendly API for creating agents with custom skills, workflows, streaming, vector databases, and state management - all with full type safety.

@ponythewhite/base-context-agent

Synerise base-context agent runtime with context management and native recovery

@xyne/workflow-sdk

Workflow engine SDK — steps, triggers, executor, agents, and a framework-agnostic HTTP router.

@llumiverse/common

Public types, enums and options used by Llumiverse API.

@b4run/langchain

LangChain adapters for B4.run agents, chains, tools, streaming, embeddings, and retry.

@pnpm/network.agent

An http/s agent factory with proxy support

@b4run/langgraph

LangGraph.js adapters and route contracts for B4.run agents, workflows, and graphs.

@owlmeans/agent-common

Serializable contracts for OwlMeans agents and LLM pipelines: conversation identity, the run lifecycle flow, and the record shapes an application persists.

@answerline/langchain

LangChain.js tools for the AnswerLine answer and search monitoring API.

@h1deya/langchain-google-ex

Drop-in replacement for ChatGoogle that automatically fixes schema compatibility issues with MCP tools and Gemini API

@relkit/agents

RELKIT agent authoring API; prefer @relkit/app/agents in applications.

@minded-ai/mindedjs-types

Type-only contract of the MindedJS agent model (flows, nodes, tools, state). Zero runtime dependencies.

langchain-cursor

LangChain chat model for Cursor models via the Cursor Agents SDK (local runtime)

@band-ai/sdk

Band TypeScript SDK core runtime

@owlmeans/llm

LLM inference runtime for OwlMeans applications: a resilient four-method model over any LangChain chat model, a provider-plugin layer, a model factory service, and an execution abstraction that resolves models from an inheritable policy.

pi-goal-x

Adds /goal to pi: conversational goal planning, flexible or ordered goals, persistent progress, and an independent completion auditor.

satohub-langchain-tools

LangChain.js tools for Sato Hub: search the scored onchain-agent index, run a Preflight check before installing or trading, get a swap-venue recommendation with the fee disclosed, and turn a goal into a build plan of real listings. Read-only and non-custo

gjallar-sdk

TypeScript/Node.js SDK for building Gjallar Protocol compatible agents — wire-aligned with the Python `gjallar` package.

@rejelly/adapter-langchain

Convert LangChain tools to Rejelly ToolDefinition for @rejelly/core

@bluebag/langchain

LangChain integration for Bluebag

@lostgradient/operative

A composable agent loop that wires Armorer and Conversationalist into the standard agentic cycle. Includes provider subpaths for Anthropic, OpenAI, and Gemini (dynamically imported — zero-SDK-if-unused).

@archstone/agent

Embed Archstone capabilities directly in your own agent loop: typed tool definitions for Anthropic, OpenAI, Gemini and plain JSON Schema, plus fail-closed execution over a compiled IR — no MCP server process required.

@tickernelz/paperclip-pro-hermes-paperclip-adapter

Paperclip adapters for Hermes Agent local CLI and Hermes Gateway HTTP/SSE runs

@tardie/ai-openai-compat

An OpenAI compat integration for Effect

@mozaik-ai/core

A TypeScript framework for building interoperable AI agents

@tardie/ai-openai

An OpenAI provider integration for Effect AI SDK

@assemblyline-agents/openai

OpenAI model provider plugin for Assembly Line's Pi harness.

@zhivex-ai/agents

Agent-first facade for the Zhivex AI SDK runtime.

@opencode/ai

Schema-first APIs for text, images, video, decisions, and embeddings, built with Effect. Speech, transcription, extraction, and realtime voice are [experimental](#experimental).

showrunner-agent-staging

Provider-neutral connector SDK for Showrunner agents

@jasonneo99/agent-workflow

Portable, model-agnostic agent workflows for any codebase

@openshain/agent

Tool loop and model providers (bring your own key)

@vinhnt-sdk/provider-openai-compatible

OpenAI-compatible model provider - raw fetch + SSE streaming, zero AI SDK, retry/backoff with ERR_UPSTREAM_* errors

@vesk/agentic

Vesk agentic plugin — zero-deps agent loop, provider abstraction, permissions, tools, checkpoints

@sogni-ai/sogni-protocol

Language-neutral protocol artifacts for the Sogni ecosystem: tool schemas, prompts, OpenAI tool manifests, and enums. Consumed by every Sogni SDK (TypeScript, Swift, and future Python/Kotlin/Rust SDKs) so contracts stay in lockstep across languages.

twilio-agent-connect

Twilio Agent Connect - A TypeScript framework for building intelligent agents

@claudebuddy/claudebuddy-agent-sdk

Open-source Agent SDK. Runs the full agent loop in-process — no local CLI required. Deploy anywhere: cloud, serverless, Docker, CI/CD.

@cubicecho/agent-core

The endpoint-agnostic half of an OpenAI-compatible agent loop: tool-schema compatibility, on-demand tool loading, one-shot side tasks, run events, and a pooled client.

@oharness/engine

The agent loop, tool execution, context management and permissions. No frontend, no transport.

@zenning/ai

AI SDK by Vercel - The AI Toolkit for TypeScript and JavaScript

@stratusagent/provider-codex

ChatGPT/Codex runtime for Stratus Agent: runs turns through the OpenAI Codex harness with kernel tools served over a loopback MCP endpoint

@pushary/openai-agents

Phone approvals for OpenAI Agents SDK agents. Your agent asks, your user taps Approve or Deny.

@yachiyo-5i/xlyra-agent

通用对话式 Agent 执行引擎:OpenAI Responses / Anthropic Messages 双协议 + Agent loop + 文件工具 + 可选 REST/SSE 服务层

@juzser/blacksmith

Autonomous agent factory — loop runner and worktree engine.

@stateofintent/agent-sdk

Client library for participating in the Epistemic Resonance Protocol as an autonomous agent

@putervision/agent-reasoning-mcp

Strategic BDI reasoning engine for autonomous AI agents — goal decomposition, utility scoring, risk evaluation, and reactive replanning over PuterVision memory and world models.

@wix/agent-platform-bm-sdk

Browser integration hooks for AI Assistant side effects. Host applications use this library to react to events emitted by assistant tools and widgets. It works alongside the Chat Service SDK, which supplies the event connection.

@deepseek-ai/dsh-session-format-v1-to-v2

Frozen released-v1 Session codec and assistant-stream migration to v2

@ai-sdk/typesafe-ai

AI SDK integration for Typesafe AI

@typesafe-ai/sdk

TypeScript SDK for the TypeSafe API

@deepseek-ai/dsh-acp

Automation-only Agent Client Protocol server for driving DeepSeek Harness agents over JSON-RPC stdio

@deepseek-ai/dsh-sdk-minimal

The standalone minimal SDK profile bundle: JSON-RPC, one DeepSeek adapter, persistent shell, editor, and JSONL sessions

@athenaintel/react

Athena React SDK — Build AI-powered chat applications with the Athena platform

@deepseek-ai/dsh-sdk-client

TypeScript client SDK for driving a DeepSeek Harness runtime subprocess over stdio JSON-RPC: the DeepSeekHarness high-level turns API and the lower-level HarnessClient

@aglyn/plugins-ai

The AI plugin for Aglyn: the in-console assistant, the Besigner copy assistant, generative building and automation jobs, and the usage metering and access controls around them. It is provider-generic: every feature talks to a provider contract, and vendor

More in Agent frameworks & infrastructure