Poe server bot quick start

Welcome to the Poe server bot quick start.

This repository serves as a companion to our tutorial and contains starter code that allows you to quickly get a bot running. The following are some of the examples included in this repo.

Note that the starter code assumes you have Modal setup for deployment (the instructions for which are described in the aforementioned tutorial).

EchoBot

• This bot simply repeats the user's query in the response.
• Before you build any server bots, you should start with reproducing this bot.
• This will ensure that your have a working fastapi_poe and modal setup.
• To deploy, run modal deploy echobot.py

A correct implementation would look like https://poe.com/EchoBotDemo

PromptBot

• This bot is an implementation of the prompt bot as a server bot.
• It demonstrates how to use the Poe platform to cover the inference costs for your chatbot.
• This bot uses Claude-3-Haiku and the system prompt instructs the bot to produce Haikus.
• If you intend to call Poe server to build your bot response, you should check if you can reproduce this bot.
• To deploy, run modal deploy prompt_bot.py
• Before you are able to use the bot, you also need to synchronize the bot's settings with the Poe Platform, the instructions for which are specified here.

A correct implementation would look like https://poe.com/PromptBotDemo

WrapperBot

• This bot is an implementation of the prompt bot as a server bot, but your own model provider API key.
• It demostrates how to wrap OpenAI API.
• You will need your OpenAI API key.
• To deploy, run modal deploy wrapper_bot.py

A correct implementation would look like https://poe.com/WrapperBotDemo

SDXLBot

• This bot demonstrates how to wrap an image generation endpoint.
• It uses Fireworks AI's StableDiffusionXL endpoint. You will need to provide your own key for this.
• Alternatively, you can change the URL to point at a different provider or model.
• To deploy, run modal deploy sdxl_bot.py

A correct implementation would look like https://poe.com/StableDiffusionXL

CatBot

• A sample bot that demonstrates the Markdown capabilities of the Poe API.
• See instructions here.
• To deploy, run modal deploy catbot.py

A correct implementation would look like https://poe.com/CatBotDemo

ImageResponseBot

• A bot that demonstrates how to render an image in the response using Markdown.
• To deploy, run modal deploy image_response_bot.py

A correct implementation would look like https://poe.com/ImageResponseBotDemo

VideoBot

• A bot that demonstrates how to attach files to your bot response. This example specifically uses video, but outputting other file types is fairly similar.
• To deploy, run modal deploy video_bot.py
• Before you are able to use the bot, you also need to synchronize the bot's settings with the Poe Platform, the instructions for which are specified here.

A correct implementation would look like https://poe.com/VideoBotDemo

PDFCounterBot

• A bot that demonstrates how to enable file upload for the users of your bot.
• To deploy, run modal deploy pdf_counter_bot.py
• Before you are able to use the bot, you also need to synchronize the bot's settings with the Poe Platform, the instructions for which are specified here.

A correct implementation would look like https://poe.com/PDFCounterBotDemo

FunctionCallingBot

• A bot that demonstrates how to use the Poe API for function calling.
• To deploy, run modal deploy function_calling_bot.py
• Before you are able to use the bot, you also need to synchronize the bot's settings with the Poe Platform, the instructions for which are specified here.

A correct implementation would look like https://poe.com/FunctionCallingDemo

FunctionCallingLoopBot

• A bot that manages a function calling loop and preserves function (tool) details in the chat for later user queries.
• To deploy, run modal deploy function_calling_loop_bot.py
• Before you are able to use the bot, you also need to synchronize the bot's settings with the Poe Platform, the instructions for which are specified here.

A correct implementation would look like https://poe.com/FunctionCallingLoopDemo

LogBot

• Illustrate what is contained in the QueryRequest object.
• To deploy, run modal deploy log_bot.py

A correct implementation would look like https://poe.com/LogBotDemo

HTTPRequestBot

• Provides an example of how to access HTTP request information in your bot.
• To deploy, run modal deploy http_request_bot.py

A correct implementation would look like https://poe.com/HTTPRequestBotDemo

TurboAllCapsBot

• This bot responds to the user's query using GPT-3.5-Turbo.
• It demonstrates how to use the Poe platform to cover the inference costs for your chatbot.
• To deploy, run modal deploy turbo_allcapsbot.py.
• Before you are able to use the bot, you also need to synchronize the bot's settings with the Poe Platform, the instructions for which are specified here.

A correct implementation would look like https://poe.com/AllCapsBotDemo

TurboVsClaudeBot

• This is a more advanced example that demonstrates how to render output in realtime comparing two different bots.
• To deploy, run modal deploy turbo_vs_claude.py
• Before you are able to use the bot, you also need to synchronize the bot's settings with the Poe Platform, the instructions for which are specified here.

A correct implementation would look like https://poe.com/TurboVsClaudeBotDemo

ClothingRecommenderBot

• An example bot that takes in a user-uploaded image, analyzes it with Claude-Sonnet-4.6 to recommend a new top, and then generates an image of the new clothing with Imagen-4 to return to the user.
• This is a good starting point for handling user-uploaded attachments, and also returning attachments with your bot.
• To deploy, run modal deploy new_top_recommender.py

A correct implementation would look like https://poe.com/TopRecommender

PythonRunnerBot

• An example bot that generates code based on the user query, runs it with the @Python bot, and attempts to debug it if there's any issue.
• This is a good starting point for chaining requests to various text models.
• To deploy, run modal deploy python_runner.py

A correct implementation would look like https://poe.com/PythonCodeRunner