imperal-sdk 5.9.3

SDK for building Imperal Cloud extensions

Imperal SDK

Build extensions for Webbee 🐝 — the agent of Imperal Cloud, the world's first AI Cloud OS.

Write a small Python function. Webbee calls it when a user asks for it — in their own words. Ship it to the Marketplace and get paid.

Documentation · Quickstart · PyPI · imperal.io

---

What this is

Imperal Cloud is the world's first AI Cloud OS — a cloud you connect the contexts of your life into (mail, money, projects, servers, notes, anything) and then run entirely in your own native language. Webbee 🐝 is the AI agent that lives inside it and does the work for you, safely.

The Imperal SDK is how you give Webbee a new skill. You write a small, typed Python extension; Webbee picks it up and calls it whenever a user asks for what it does — in plain language. The platform handles the hard parts — authentication, billing, multi-tenancy, audit, recovery, LLM routing — so your code stays small.

pip install imperal-sdk

Python 3.11+ · Apache-2.0

Why build here

• Webbee calls your function directly. Typed, structured calls — no LLM guessing your arguments, no silent write failures.
• The platform does the plumbing. Auth, per-user isolation, billing, audit, retries and recovery, multi-tenant safety — handled for you.
• You get paid. Publish to the Imperal Marketplace, price your extension, earn on every use.
• Bring any LLM. Users connect their own model keys — Anthropic, OpenAI, Google, Ollama, any OpenAI-compatible API.

A 60-second extension

from imperal_sdk import Extension, ChatExtension, ActionResult
from pydantic import BaseModel, Field

ext = Extension(
    "hello-world",
    version="1.0.0",
    display_name="Hello World",
    description="Greets people by name with a friendly message.",
    icon="icon.svg",
    actions_explicit=True,
)

chat = ChatExtension(ext, tool_name="hello_world", description="Friendly greetings.")


class GreetParams(BaseModel):
    name: str = Field(..., description="Person to greet")


@chat.function("greet", description="Greet someone by name.", action_type="read")
async def greet(ctx, params: GreetParams) -> ActionResult:
    return ActionResult.success(
        data={"message": f"Hello, {params.name}! 🐝"},
        summary=f"Greeted {params.name}",
    )

That's a real, working extension. When a user types "say hi to Alex", Webbee calls greet(name="Alex").

→ Full walkthrough, from zero to published: docs.imperal.io

What you can build

• Chat tools — typed @chat.functions Webbee calls straight from natural language.
• Panels — UI surfaces rendered inside the Imperal Panel.
• Skeletons — live data feeds that keep Webbee aware of a user's state.
• Scheduled jobs & webhooks — act on a timer, or react to outside events.
• Typed entities (SDL) — return sdl.Entity objects and the platform reads their meaning (id, title, kind, …) directly instead of guessing field names. Live in production.

Every published extension passes the federal contract — the validators that let Webbee call your code safely. The SDK checks it locally before you ship.

Test without a server

from imperal_sdk.testing import MockContext

async def test_greet():
    ctx = MockContext()
    result = await greet(ctx, GreetParams(name="Alex"))
    assert result.status == "success"

Documentation

The full API, the manifest schema, every validator, recipes, and the federal contract live at docs.imperal.io — that's the source of truth; this README is just the doorway.

Documentation	docs.imperal.io
Quickstart	docs.imperal.io/en/getting-started/quick-start
PyPI	pypi.org/project/imperal-sdk
Changelog	CHANGELOG.md
License	Apache-2.0

---

Built by Imperal, Inc. — Webbee 🐝 is its agent.