ListFrom Stakk
Frameworks and platforms for building AI agents
From TypeScript or Python code to visual canvases: eight tools for building agents with memory and RAG. What each one is for.
8 resourcesUpdated Oct 3, 2026
An AI agent combines a model, tools it can call and memory of what already happened. These tools handle those pieces in different ways: some are libraries you code against, some are visual canvases, and one covers a single piece, memory.
Pick by your team's language first. If you work in TypeScript, start with the AI SDK or Mastra; in Python, with LangGraph or CrewAI. If nobody codes, look at Dify, Flowise or Botpress.
Most are open source and free as libraries. What you pay for is the managed platform and, always, model usage.
AI SDKFree library to call dozens of model providers through one API, with streaming, tool calls and structured output. Memory and durable workflows are up to you.
MastraTypeScript framework that ships with memory, pausable and resumable workflows, RAG, evals and MCP support. A natural fit for JavaScript teams that don't want to switch to Python.
LangGraphModels the agent as a stateful graph, with durable execution and pauses so a person can approve a step. You get fine-grained control in exchange for a steep learning curve.
CrewAIOrganizes teams of agents with a role, a goal and tools, plus “flows” to control execution. It's Python only and doesn't depend on LangChain.
DifyVisual canvas for chatbots, agents and RAG workflows, published as an API or web chat. The interface is available in Spanish; use their cloud or self-host with Docker.
FlowiseBuild agents and multi-agent flows by dragging nodes on top of LangChain. It supports local models like Ollama if you'd rather keep data in-house.
BotpressAI agents for customer service that publish to WhatsApp, Instagram and your site chat, and take real actions once you connect your APIs. No per-seat pricing.
Mem0Memory layer that pulls the key facts out of each conversation and retrieves them later, so you don't resend the full history to the model. Works with the main vector databases and frameworks.
