TL;DR: AI in Beeper Chats is a feature on Beeper Desktop that lets you tag @ChatGPT or @Apple Intelligence directly in any conversation to ask questions and get AI-powered responses, all without leaving the chat. It requires a Beeper Plus or Plus Plus subscription. Responses are sent from your device and visible to everyone in the chat (consider this when asking your AI in group chats). The AI can reference recent conversation history, and you can customize its behavior with a custom system prompt in the app settings.

What is it?

AI in Beeper Chats is a feature on Beeper Desktop that lets you mention and invoke AI assistants, @ChatGPT and @Apple Intelligence, directly within any chat conversation. By tagging one of these AI models in a message, you can ask questions, get help, or generate responses without leaving the chat. The AI's reply appears in the same conversation thread as a message sent from your device, visible to all participants.

You can also customize the AI's behavior by setting a Custom System Prompt in Settings (e.g., giving it a specific persona or set of instructions). The AI has context awareness and can reference roughly the last 20 messages in the conversation to give relevant answers. Additionally, ChatGPT can describe images when you use the Reply function on an image message (Apple Intelligence has limited image capabilities).

Who has access to it?

Access requires all the following:

• Beeper Plus or Plus Plus subscription
• Beeper Desktop app (macOS, Windows, or Linux) — the feature is not yet available on Beeper mobile apps.
• For Apple Intelligence specifically: macOS 26 (Tahoe) or later and Apple Silicon (M1 chip or newer) are required. Apple Intelligence is not available on Windows or Linux.

Users must opt in by going to Settings → General and enabling “Mention an AI to get answers and help in chat.”

How does it work?

1. Enable the feature: Go to Settings → General and toggle on "Mention an AI to get answers and help in chat."
2. Mention the AI: In any DM or group chat, type @ to see the autocomplete list, then select @ChatGPT or @Apple Intelligence. Type your question or request alongside the mention (e.g., "@ChatGPT What is the capital of France?").
3. First-time use: On a brand-new account with the setting still off, attempting to type an AI mention triggers a warning modal explaining the feature. Confirming the modal allows the mention; cancelling removes it.
4. Thinking phase: After sending, a placeholder message appears (e.g., "ChatGPT is thinking..." or "Sent by ChatGPT on Beeper") while the AI processes the request.
5. Response delivery: After a few seconds, the placeholder updates in-place with the AI's actual response. The message is visually distinguished as an AI response (e.g., Sent by ChatGPT on Beeper). The response is sent from your device and is visible to all members in the chat.
6. Context & images: The AI can reference recent conversation history (approximately the last 20 messages). ChatGPT can describe images when you reply to an image message; it cannot see images sent directly in the same message. Apple Intelligence has limited/vague image description capability.
7. Custom System Prompt: In Settings → General, you can enter a custom prompt to change the AI's personality or behavior. Clearing the prompt resets it to the default helpful assistant tone.
8. Error handling: If something goes wrong (e.g., you lose internet), the "thinking" placeholder changes to "Generation Failed," a toast notification appears with error details, and a Retry button allows up to 3 retry attempts before it is disabled.