Agent Frameworks

20+ frameworks for building AI agents in TypeScript and Python

20
Total Frameworks
10
TypeScript
18
Python
Language:
Category:
Showing 20 of 20 frameworks

LangChain

LangChainBoth

Full-stack framework for building LLM-powered applications with composable chains, agents, memory, and retrieval.

130k+
Composable chains and agentsBuilt-in memory systemsRAG and retrieval pipelinesTool calling and function execution+4 more

Best for: Teams needing a batteries-included framework with broad LLM provider support and production tracing.

LangGraph

LangChainBoth

Library for building stateful, multi-agent applications as cyclic graphs with built-in persistence and human-in-the-loop support.

25k+
Stateful graph-based agent flowsCyclic and branching executionBuilt-in persistence and checkpointingHuman-in-the-loop breakpoints+4 more

Best for: Complex multi-agent workflows that need state management, branching logic, and production deployment.

CrewAI

CrewAIPython

Role-based multi-agent orchestration framework where AI agents collaborate as a crew with defined roles, goals, and backstories.

45k+
Role-based agent designSequential and parallel task executionBuilt-in tool ecosystemAgent delegation and collaboration+4 more

Best for: Teams wanting an intuitive role-based approach to multi-agent systems with minimal boilerplate.

OpenAI Agents SDK

OpenAIPython

Official OpenAI framework for building agentic applications with handoffs, guardrails, and tracing built in.

18k+
Agent handoffs between specialistsInput and output guardrailsBuilt-in tracing and observabilityTool calling with type safety+4 more

Best for: OpenAI-centric teams wanting first-party agent primitives with handoffs and safety guardrails.

Anthropic SDK

AnthropicBoth

Official Anthropic SDK for building agents with Claude, featuring tool use, multi-turn conversations, and extended thinking.

3k+
Tool use and function callingExtended thinking for complex reasoningMulti-turn agent loopsStreaming with tool use+4 more

Best for: Teams building with Claude who want direct SDK access to tool use, extended thinking, and computer use capabilities.

Vercel AI SDK

VercelTypeScript

Full-stack TypeScript toolkit for building AI applications with streaming, tool calling, multi-step agents, and structured output.

22k+
Streaming UI with React Server ComponentsMulti-step tool calling agentsStructured output with Zod schemasMulti-provider support (OpenAI, Anthropic, Google, etc.)+4 more

Best for: TypeScript/Next.js developers building full-stack AI apps with streaming UIs and multi-provider support.

AutoGen

MicrosoftPython

Microsoft framework for building multi-agent conversational systems where agents can chat with each other to solve tasks.

55k+
Multi-agent conversationsCode generation and executionHuman-in-the-loop integrationFlexible agent topologies+4 more

Best for: Research teams and enterprises needing flexible multi-agent conversations with code execution capabilities.

Composio

ComposioBoth

Tool integration platform that gives AI agents access to 250+ external tools and services with managed auth and execution.

27k+
250+ pre-built tool integrationsManaged OAuth and authenticationWorks with any agent frameworkGitHub, Slack, Gmail, Jira, and more+4 more

Best for: Teams that need agents to interact with real-world SaaS tools without building individual integrations.

Mastra

MastraTypeScript

TypeScript-first agent framework with built-in workflows, RAG, integrations, and an evaluation system for AI applications.

22k+
TypeScript-first designDeclarative workflow engineBuilt-in RAG with vector storesAgent memory and context+4 more

Best for: TypeScript developers wanting an opinionated, batteries-included agent framework with workflows and RAG.

Pydantic AI

PydanticPython

Type-safe Python agent framework built on Pydantic, bringing validation, structured outputs, and dependency injection to AI agents.

15k+
Pydantic-based structured outputsType-safe tool definitionsDependency injection for toolsMulti-model support+4 more

Best for: Python developers who value type safety and want Pydantic's validation power applied to agent development.

BeeAI Framework

IBMBoth

Open-source framework for building production-ready AI agents in Python and TypeScript with tool use and memory.

3k+
Python and TypeScript agentsReAct agent patternBuilt-in tool libraryMemory and context management+4 more

Best for: Developers wanting a lightweight, IBM-backed agent framework with production observability.

smolagents

Hugging FacePython

Minimal Python agent framework by Hugging Face focused on code-based agents that write and execute Python to solve tasks.

25k+
Code-based agent executionMinimal API surfaceMulti-step reasoningTool calling via code generation+4 more

Best for: Developers wanting a simple, code-first agent framework that leverages the Hugging Face ecosystem.

Semantic Kernel

MicrosoftBoth

Microsoft enterprise AI orchestration SDK for integrating LLMs into conventional applications with plugins and planners.

27k+
Enterprise-grade AI orchestrationPlugin architecture for toolsPlanner for multi-step executionMemory and vector store integration+4 more

Best for: Enterprise teams integrating AI into existing .NET or Python applications with governance requirements.

Haystack

deepsetPython

Modular NLP framework for building production-ready RAG pipelines, search systems, and question-answering applications.

24k+
Modular pipeline architectureRAG pipeline buildingDocument stores and retrieversCustom component creation+4 more

Best for: Teams building production RAG systems and search pipelines with a modular, composable architecture.

LlamaIndex

LlamaIndexBoth

Data framework for connecting LLMs to external data sources with advanced indexing, retrieval, and query engine capabilities.

47k+
Advanced data indexing and retrievalQuery engine abstractionAgent and tool abstractionsMulti-document agents+4 more

Best for: Teams building data-intensive LLM applications that need sophisticated retrieval and indexing over private data.

Agno

AgnoPython

Full-stack AI agent framework (formerly PhiData) for building agents with memory, tools, knowledge bases, and multi-agent teams.

38k+
Built-in memory and knowledge basesMulti-agent teams with coordinationTool calling with pre-built toolkitsStructured outputs and validation+4 more

Best for: Python developers wanting a batteries-included agent framework with memory, knowledge, and team coordination out of the box.

Agency Swarm

VRSENPython

Multi-agent orchestration framework for creating collaborative AI agent swarms with customizable roles and communication flows.

4k+
Customizable agent roles and instructionsInter-agent communication flowsOpenAI Assistants API integrationShared state management+4 more

Best for: Developers building collaborative multi-agent systems with defined communication hierarchies and shared state.

Instructor

Jason LiuBoth

Lightweight library for extracting structured outputs from LLMs using Pydantic models, with retries, validation, and streaming support.

12k+
Structured output extraction with PydanticAutomatic retries with validationStreaming partial responsesMulti-provider support (OpenAI, Anthropic, etc.)+4 more

Best for: Teams that need reliable structured data extraction from LLMs with type-safe validation and retry logic.

DSPy

Stanford NLPPython

Stanford framework for programming - not prompting - language models using composable modules, optimizers, and automated prompt tuning.

32k+
Declarative language model programmingAutomatic prompt optimizationComposable modules (ChainOfThought, ReAct, etc.)Built-in evaluation and metrics+4 more

Best for: Research teams and ML engineers who want to optimize LLM pipelines programmatically instead of hand-tuning prompts.

Outlines

.txtPython

Structured text generation library that guarantees LLM outputs match a given format using regex patterns, JSON schemas, or grammars.

13k+
Regex-guided text generationJSON schema-constrained outputContext-free grammar supportMultiple choice generation+4 more

Best for: Teams running local models that need guaranteed structured output conforming to schemas, regex, or grammars.