AICHE +

P

Poe Integration

The short answer: open Poe, select your preferred AI model, click into the chat field, press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux), speak your prompt for 30-60 seconds, and AICHE inserts the formatted query for any model on the platform.

The Model Comparison Problem

Poe by Quora gives you access to GPT-4, Claude, Gemini, Llama, Mistral, and dozens of other models from one interface. The whole point is comparing how different models handle the same question.

But here is what actually happens: you type a detailed prompt to one model, read the response, switch to another model, and then need to type the same prompt again. Or you copy-paste, which means you need to have composed the detailed prompt in the first place. Most people end up asking a simpler version of their question because sending a thorough prompt to three models means typing it three times or managing clipboard pasting.

With AICHE, you dictate once. The transcribed text sits in the input field. You send it to the first model, copy the text, switch models, paste, send. Or open three Poe tabs, dictate once, paste into the other two. The voice transcription creates a reusable prompt that works identically across every model.

How to Use It

1. Open Poe in your browser or desktop app (poe.com).
2. Select an AI model from the sidebar (Claude, GPT-4, Gemini, or any available model).
3. Click into the chat input field.
4. Press your AICHE hotkey (⌃+⌥+R on Mac, Ctrl+Alt+R on Windows/Linux) to start recording.
5. Speak your complete question with all relevant detail (example: "explain the trade-offs between using WebSockets versus Server-Sent Events for real-time updates in a React application, covering browser compatibility, server resource usage at 10,000 concurrent connections, reconnection strategies, and include basic implementation patterns for both approaches").
6. Press the hotkey again. AICHE transcribes, applies Message Ready formatting, and inserts the text.
7. Send to the current model, then switch models and use the same prompt to compare responses.

Using Custom Bots

Poe lets you create and use custom bots, which are models with specific system prompts and configurations. When chatting with a custom bot, the prompt you send still matters. The bot's system prompt shapes its personality and focus, but your message determines what it actually works on.

Dictating detailed prompts to custom bots is especially valuable because these bots are purpose-built. A custom bot for code review needs to know the language, framework, and what aspects of the code concern you. A custom bot for writing needs to know the audience, tone, and format. The more detail you give it, the better it leverages its specialized configuration.

Evaluating Model Strengths

When you are comparing models, consistency matters. The same prompt sent to different models lets you compare apples to apples. If you shorten or rephrase the prompt between models, you are testing prompt quality, not model quality.

Voice transcription gives you a consistent, detailed prompt every time. Dictate once, reuse the exact text. This makes Poe's multi-model access genuinely useful for evaluation instead of just convenient for chatting.

Pay attention to how models differ on the same input. Some handle ambiguity better. Some give longer, more structured responses. Some are better at code. Having a consistent detailed prompt makes these differences visible.

Heads-up: Poe shows conversation history per model. If you want clean comparisons, start new conversations with each model rather than continuing existing threads that might carry different context.

Pro tip: enable Clean Language in AICHE settings. Your spoken questions get polished into clear, well-structured prompts that work well across all models. Different models handle ambiguous prompts differently, but clean language improves responses from all of them.

Result: sending the same technical question to three different models for comparison, which took 18 minutes with typing, now takes 45 seconds of dictation plus 30 seconds of copy-paste. You get real comparisons instead of settling for one model's answer.

Do this now: open Poe, select two different models, press your hotkey, and ask one technical question to both. Compare the reasoning, depth, and structure of each response.