Image Upload

How to Use

Click the 🖼️ button to upload, or use Cmd+V / Ctrl+V to paste an image from the clipboard.

Supported Formats

JPEG, PNG, GIF, WebP.

Image Preview

After uploading, a thumbnail preview is displayed. You can remove individual images, and attaching multiple images at once is supported.

Supported Models

Provider / Type	Description
OpenAI	GPT-4o series and other vision-capable models
Google	Gemini 2.5, etc.
Alibaba	Qwen VL
OpenRouter	Depends on whether the selected model supports vision
Ollama	LLaVA and other multimodal local models

⚠️ Model Capability Models without vision support may not process image content correctly.

Images in AI Replies

In addition to uploading images to the AI, Momo can also display images returned in AI replies (common with Hermes, OpenClaw, and other agents that generate images).

When a reply contains any of the following, Momo tries to download and render the image inline:

A MEDIA:<https-url> marker
Markdown image syntax ![...](https://...)
HTTP/HTTPS URLs ending in .png, .jpg, .jpeg, .webp, or .gif

Successfully downloaded images are stored locally in IndexedDB (AttachmentStore) and rendered in the chat bubble, not just as external links. OpenClaw replies that include HTTP images via image / image_url content parts follow the same flow.

⚠️ Download limits If a URL has expired, requires sign-in, or is not reachable over HTTP/HTTPS, Momo may fail to download it and you may still see a link or blank area. This is an interim solution; behavior may change when agents provide a dedicated media endpoint.

Technical Details (Upload)

Images are sent as Base64-encoded data. The request format follows the OpenAI Vision API convention using image_url content blocks.