Omni Judge · Tech Demos · theserverless.dev

What this proves Clef-omni reads the media and scores every allowed answer in one request. The bundled presets are real intake tasks with a labelled expected answer — no transcribe-then-classify chain unless you ask for the comparison.

What to try

Run a preset: expense receipt, product QA, delivery parcel, support voice, or a safety clip.
Read the top answer, the bar for every option, and whether it matched the expected label.
Snap a photo, record a voice note, or film up to 10 seconds and write your own checklist.
Compare an image or a voice note with Whisper or a vision model, then Clef.

What you could build

Expense intake that rejects a blurry receipt.
Product QA that flags a cracked mug or phone screen.
A delivery photo check for a damaged parcel.
Support voice-note triage, urgent versus routine.
A short safety clip that spots a fall.

Media you upload is deleted after 24 hours. Model text is untrusted.

Media

Choose a file Use the camera roll

On an iPhone, the camera roll can return HEVC or a file over the size limit. Photos are resized here. Video must be MP4 or WebM, within 10 seconds and 6 MB.

Nothing selected yet.

Checklist

Also run the classic chain and compare latency

On a video, the chain is skipped. Whisper and the vision model do not read the clip.

Recent judgments

Judgments from this browser show up here for 24 hours.