Langflow
Free tierLow-code AI builder for agentic and RAG applications
Free tier available·All audiences·Powered by Multi-provider (supports all major LLMs including Anthropic, OpenAI, Meta, Mistral, NVIDIA, Ollama, Groq, HuggingFace, Amazon Bedrock, Perplexity, and more)·API available·Open source
Key strengths
Visual low-code flow builder with drag-and-drop interfaceSupports all major LLMs and vector databasesPython customization under the hoodAgentic and RAG application supportFree cloud deployment and self-hosting optionsLarge library of pre-built flows and componentsMCP server supportOpen source (MIT License)
Free tier + paid plans
Founded 2023
Self-hostable
No ratings yet
Langflow is purpose-built for engineering teams developing:
- RAG pipelines: Compose end-to-end retrieval-augmented generation workflows by connecting document loaders (Unstructured, Airbyte, Google Drive, Confluence), vector stores (Pinecone, Weaviate, Milvus, Qdrant, Cassandra), and LLMs in a single visual flow.
- Agentic applications: Build single-agent or multi-agent systems with tool use, leveraging native CrewAI support and integrations with Composio, Tavily, Serper, Wolfram Alpha, and other tool providers.
- LLM-powered APIs: Wrap any Langflow pipeline as an API endpoint and embed it into existing products or backend services.
- Multi-model experimentation: Rapidly swap and benchmark LLM providers (OpenAI, Anthropic, Groq, Mistral, Ollama, etc.) within the same flow architecture without rewriting pipeline logic.
- Enterprise data integrations: Connect internal knowledge sources (Confluence, Notion, GitHub, Google Drive, MongoDB) to LLM workflows for internal search, Q&A, and automation use cases.
- MCP-compatible tooling: Integrate Langflow as an MCP server within broader model-tool ecosystems.