AI agent SDKs, platforms & frameworks
The agent layer of the AI stack — from SDKs you ship in code (OpenAI, Claude) to no-code platforms (Lindy, Dify) and orchestration frameworks (LangChain, CrewAI). Each review covers pricing, lock-in, and what you should pair it with.
Choose the model first
Decide whether the work needs an agent before choosing a framework
Start with the execution boundary, then narrow the framework and deployment choice.
- Check the execution model Use predictable workflows for explicit steps, agents for bounded judgment, and durable execution when state must survive failures or waits. Review the three models →
- Confirm an agent fits Choose by workload shape, state, provider boundary, deployment ownership, and who maintains the system. Read the framework decision guide →
- Compare the shortlisted options Use head-to-head evidence once the execution model and operating constraints are clear. Open comparisons →
AutoGen
Microsoft's open-source multi-agent framework — conversation-driven orchestration with deep Azure / OpenAI integration.
Claude Agent SDK
Build agents with tool use, computer use, MCP, and Anthropic's safety primitives — Python and TypeScript.
CrewAI
OSS framework for orchestrating role-playing AI agents that collaborate on complex tasks — Python-first.
Dify
Self-hostable platform for agentic AI apps — RAG pipelines, agent workflows, and model management in one stack.
Flowise
Apache-licensed visual builder for LLM agents and chatflows — self-hostable, with upstream end-of-life announced for August 2026.
LangChain
Open-source framework ecosystem for LLM applications, with LangGraph for stateful orchestration.
Langflow
Open-source visual builder for LangChain-style agents and RAG flows — drag-and-drop, Python-native, self-hostable.
LangGraph
Low-level graph-based agent orchestration from the LangChain team — explicit state, loops, and human-in-the-loop control.
Lindy
No-code AI assistants for sales, support, and operations workflows across email, meetings, and calls.
OpenAI Agents SDK
Lightweight framework for multi-agent workflows with handoffs, guardrails, and tool calls — Python and TypeScript.