Skip to main content

​
What is floom?

floom is the production layer for AI agents. You write a Python function, and floom turns it into:
  • A web UI with typed inputs and outputs anyone can use
  • A REST API other services can call
  • An MCP endpoint AI agents can discover
  • Managed secrets, version history, and scheduling
No Docker. No CI/CD. No YAML. Just Python in, production out.

​
How it works

Write a run() function → Deploy via AI agent or API → Share a live URL
  1. Write a Python script with a run() function
  2. Deploy via Claude Code (/floom), Cursor, or any MCP-capable agent
  3. floom tests it in an E2B sandbox, then deploys it
  4. You get a shareable URL at dashboard.floom.dev/a/<id>

​
Who it’s for

  • Developers who build Python automations and need to share them with non-technical people
  • AI agencies deploying tools for clients
  • Anyone using Claude Code or Cursor who wants to go from localhost to production instantly

​
Example

import os
from google import genai

def run(url: str) -> dict:
    client = genai.Client(api_key=os.environ["GEMINI_API_KEY"])
    response = client.models.generate_content(
        model="gemini-2.0-flash",
        contents=f"Summarize the company at {url} in 3 sentences."
    )
    return {"summary": response.text}
That’s it. floom handles sandbox testing, deployment, UI generation, and secret injection.

​
Next steps

Quickstart

Deploy your first app in 60 seconds

Manifest format

Inputs, outputs, dependencies, secrets

API reference

Authenticate and call the REST API

Templates

Starter scripts to fork and deploy