Integrate Custom LLM App into Slack using Runbear

Previously, integrating OpenAI Assistants with Runbear didn't require coding, but flexibility was limited. Now, Runbear supports custom LLM apps with the launch of the Runbear Python SDK. This means you can use any AI model or API - OpenAI, Anthropic, Gemini, and more. Let's dive into how to integrate a custom LLM app with Runbear.

Step 1: Integrate Runbear SDK into Your LLM App

Install Runbear SDK

If you're using FastAPI, use the command below to install the SDK.

pip install 'plugbear[fastapi]'

For other setups, simply use the command below.

pip install plugbear

Get Runbear API Key

You can get your API key from the Profile menu on the top right. Click "Manage API Keys" and then "New API Key". Copy the generated key.

Initialize Runbear SDK and Set Up Endpoint

Here's an example of how to initialize the SDK and set up the endpoint path that Runbear will send requests to. Replace the PLUGBEAR_API_KEY with your copied API key. You can also change the endpoint path to whatever you want.

import contextlib
from openai import OpenAI
from fastapi import FastAPI
import plugbear.fastapi

# Replace with your copied Runbear API Key
PLUGBEAR_API_KEY = "PLUGBEAR_API_KEY"

# Endpoint path for Runbear requests
PLUGBEAR_ENDPOINT = "/plugbear"

@contextlib.asynccontextmanager
async def lifespan(app: FastAPI):
    await plugbear.fastapi.register(
        app,
        llm_func=some_llm,
        api_key=PLUGBEAR_API_KEY,
        endpoint=PLUGBEAR_ENDPOINT,
    )
    yield

app = FastAPI(lifespan=lifespan)
client = OpenAI()

async def openai(context: plugbear.fastapi.Request) -> str:
    chat_completion = client.chat.completions.create(
        model='gpt-4-1106-preview',
        messages=[{"role": msg.role, "content": msg.content}
                  for msg in context.messages],
    )
    return chat_completion.choices[0].message.content

For more examples, refer to our examples on GitHub