Sidebar Chat
The core feature of Momo AI is the sidebar chat interface built with the Chrome Side Panel API.
Interface Overview
- Top toolbar: System prompt selector, chat history, new conversation, settings
- Message area: Markdown rendering, collapsible thinking process, action buttons
- Input area: Model selector, feature button bar (📄🖼️🔍💡), text input box, character counter, send button
Chat Controls
- Enter: Send message
- Cmd+Enter (macOS): Also sends message
- Shift+Enter: New line (without sending)
- Escape: Stop generation
Auto-Scroll
The conversation automatically follows the latest content. If you manually scroll up to view earlier messages, auto-scroll will pause. Click the ⬇ button to resume following the latest content.
When an AI reply includes images (stickers or Markdown images), layout height may change after images finish loading. In auto-follow mode, Momo scrolls to the bottom again so the last text is not hidden behind the input area.
Streaming Responses
Responses are streamed via SSE (Server-Sent Events). Different update frequencies are used based on response length: short responses update approximately every 80ms, medium every 300ms, and long every 800ms, balancing smoothness and performance.
Welcome Screen & Prompt Suggestions
When no conversation has started, a welcome screen shows Prompt Suggestions for one-click starters. See System Prompts for details.
The system prompt selector in the top toolbar appears only when at least one prompt is set to visible. MoMo is active by default.
Edit Messages
Click Edit on a user message to change its text or attachments and resend.
- Standard models (OpenAI, Gemini, etc.): editing replaces that user message, removes all later local messages, and regenerates the assistant reply.
- Hermes / OpenClaw: editing appends a new user message (keeping the original and full local history) and sends it to the agent / gateway. Remote agent sessions cannot be rolled back by Momo.
- Editing is blocked while a response is streaming; stop generation first.