Get Started With Opper

1. Create an API key

Log in at platform.opper.ai and create your own API key in the top right menu. Each api key is associated with a project, you will have all calls, functions, indexes and traces associated with this project. There is a default project for each organization but it is recommended to create a new project to better being able to separate the different applications and environments.

2. Set the API key

Set the API key as the OPPER_API_KEY environment variable.

3. Install the Opper SDK

Install our SDK for your preferred language. There are currently SDKs available for Python and TypeScript. If your language is not supported, you can use the REST API directly. See the API reference for more information.

3. Make your first LLM call

Let's write a simple call that extracts room descriptions from text as structured output.

import asyncio
# It is strongly recommended to use the `AsyncOpper` client to make calls asynchronously
# but if that is not an option, there is also an `Opper` client to make calls synchronously
from opperai import AsyncOpper
# Our SDK supports Pydantic to provide structured output
from pydantic import BaseModel

# Define the output structure
class RoomDescription(BaseModel):
    room_count: int
    view: str
    bed_size: str
    hotel_name: str

async def main():
    # Don't forget to set the API key as the OPPER_API_KEY environment variable!
    opper = AsyncOpper()

    # Make a call
    result, _ = await opper.call(
        name="extractRoom",
        instructions="Extract details about the room from the provided text",
        input="The Grand Hotel offers a luxurious suite with 3 spacious rooms, each providing a breathtaking view of the ocean. The suite includes a king-sized bed, an en-suite bathroom, and a private balcony for an unforgettable stay.",
        output_type=RoomDescription,
    )

    print(result)
    # RoomDescription(room_count=3, view='ocean', bed_size='king-sized', hotel_name='The Grand Hotel')

asyncio.run(main())

import asyncio
from opperai import fn

# Our SDK supports Pydantic to provide structured output
from pydantic import BaseModel

# Define the output structure
class RoomDescription(BaseModel):
    room_count: int
    view: str
    bed_size: str
    hotel_name: str

# Don't forget to set the API key as the OPPER_API_KEY environment variable!
@fn(path="extract_room", model="openai/gpt-4o")
async def extract_room(text: str) -> RoomDescription:
    """Given a text, extract the room description"""

async def main():
    print(await extract_room("The Grand Hotel offers a luxurious suite with 3 spacious rooms, each providing a breathtaking view of the ocean. The suite includes a king-sized bed, an en-suite bathroom, and a private balcony for an unforgettable stay."))  

asyncio.run(main())

import Client from "opperai";

// Don't forget to set the API key as the OPPER_API_KEY environment variable!
const client = new Client();

const roomDescriptionSchema = {
    $schema: "https://json-schema.org/draft/2020-12/schema",
    type: "object",
    properties: {
        room_count: { type: "number" },
        view: { type: "string" },
        bed_size: { type: "string" },
        hotel_name: { type: "string" },
    },
    required: ["room_count", "view", "bed_size", "hotel_name"],
};

async function main() {
    // remember to return the json_payload from the call
    const { json_payload } = await client.call({
        name: "extractRoom",
        input: "The Grand Hotel offers a luxurious suite with 3 spacious rooms, each providing a breathtaking view of the ocean. The suite includes a king-sized bed, an en-suite bathroom, and a private balcony for an unforgettable stay.",
        output_schema: roomDescriptionSchema,
    });

    console.log(json_payload);
    // Expected output: { room_count: 3, view: 'ocean', bed_size: 'king-sized', hotel_name: 'The Grand Hotel' }
}

main();

# Make a call with input and output schema (input schema is optional)
curl -X POST "https://api.opper.ai/v1/call" \
     -H "Content-Type: application/json" \
     -H "X-OPPER-API-KEY: ${OPPER_API_KEY}" \
     -d '{
            "name": "test/extract_description",
            "instructions": "Extract details about the room from the provided text",
            "input_type": {
                "type": "object",
                "properties": {
                    "text": {
                        "type": "string"
                    }
                },
                "required": [
                    "text"
                ]
            },
            "output_type": {
                "type": "object",
                "properties": {
                    "room_count": {
                        "type": "number"
                    },
                    "view": {
                        "type": "string"
                    },
                    "bed_size": {
                        "type": "string"
                    },
                    "hotel_name": {
                        "type": "string"
                    }
                },
                "required": [
                    "room_count",
                    "view",
                    "bed_size",
                    "hotel_name"
                ]
            },
            "input": {
                "text": "The Grand Hotel offers a luxurious suite with 3 spacious rooms, each providing a breathtaking view of the ocean. The suite includes a king-sized bed, an en-suite bathroom, and a private balcony for an unforgettable stay."
            }
        }'

4. View your call in the platform

Check out your call in the traces page.