Files, images, video and music in Chat
Attach documents, images, sound and video, and ask any chat model for a picture, a video, a song or a file to download.
What it is
People attach documents, images, sound and video to a message. A model that can't see, hear or watch gets a transcript or a description instead, made once and reused. Any chat model can also make things: pictures with the organization's image model, videos with Grok Imagine, Veo or Sora, music with Google Lyria, and real files such as code, pages or CSV that open as a card with Preview, Download and Copy.
Why you want it
Work happens in files. People want to drop in a PDF, a screenshot or a meeting recording and get a useful answer, and ask for a chart or an image without switching tools, while the files stay inside your own storage.
How it works
- Images (PNG, JPEG, GIF, WebP), documents (PDF, Word, text, Markdown, CSV, JSON, YAML, HTML, logs, code), sound (MP3, WAV, M4A, OGG, FLAC, AAC, Opus) and video (MP4, MOV, WebM)
- Files are checked by their bytes, not their name; up to 10 per message, each up to 20 MiB by default
- Administrators set per model whether it sees images, hears sound and sees video, and which models transcribe sound and describe video
- Each image, video or song is its own metered request, billed to the person's current team or their own plan
- Everything is kept encrypted in Janus's file storage, visible only to its owner, and deleted with the chat
- Memory: tell the assistant to remember something and later chats use it; owner-only, encrypted, and can be switched off
See it on your own network.
The Community edition is free for up to 25 people. The 30-day Business trial unlocks every Business feature.