Langflow: agents built on a canvas
Visual LangChain: drag, connect and publish flows as API
- Video memory
- 4 GB
- Setup
- ~4 min
- Access
- port 7860
- Billing
- hourly, in BRL
Langflow turns what would be orchestration code into blocks wired on a canvas. You build the flow, test it right there and publish it as an API — without writing the glue between model, vector store and tools.
Langflow is a visual IDE for LangChain. Build RAG, agents, chatbots and pipelines by dragging blocks: LLMs (OpenAI, Anthropic, local vLLM), vector stores, tools and memory. Export each flow as a REST endpoint.
What it is for
- Prototype agents and RAG in hours, not weeks
- Publish the finished flow as an API for your application
- Swap the model in a block without redoing the logic
- Connect to your own model running on the same machine
How to deploy
- Create your account and add balance (card or Pix, no subscription).
- In the console, pick the Langflow template and a machine — the console hides the ones that do not meet the requirement.
- In about 4 minutes the setup finishes and the access address shows up in the panel, on port 7860.
Done? Just destroy the machine and billing stops with it. No contract, no minimum commitment.
FAQ
Does it need a GPU?
Only if the model runs on the same machine. Pointing at an external API, a small machine is enough.
Can I export the flow?
Yes, as JSON, so you can version it in your repository.
Is it production-ready?
It is fine for prototyping and moderate load; high volume calls for your own code.