FAQ

TheiBusiness AI Analyzer examines four kinds of content: text (pasted or from a public web page), images, audio and video.

For each one, it reports the technical facts it can observe in the file (Level 1) and a set of algorithmic indicators computed from the content itself (Level 2). AI-powered analysis through specialized providers (Level 3) is coming soon.

Current limits

No. No detector, including ours, can prove that content was created by a human or by AI. The analyzer reports indicators: observations that are more or less common in generated content.

Results use cautious wording such as No significant indicators detected, Some unusual indicators detected, Multiple indicators detected or Analysis inconclusive. We never display “100% AI”.

The only exception is provenance metadata: when a file declares itself as AI-generated (for example through C2PA Content Credentials or IPTC metadata), the report says so, and notes that the declaration was not cryptographically verified.

Never use a result on its own to accuse someone of using AI.

• Text and images are sent over HTTPS to our server, analyzed in memory and deleted as soon as the analysis finishes. We do not keep a copy.
• Audio and video are analyzed directly in your browser. The file never leaves your device.
• URLs: the page or image is downloaded to a temporary file, analyzed and deleted immediately. Any leftover temporary file is removed automatically.

Nothing is sent to a third-party AI service. If AI-powered providers are added in the future, each one will be clearly identified before you send content to it.

These limits are read live from our configuration:

Audio and video decoding happens in your browser, so a format your browser cannot decode (for example some MOV codecs in Firefox) may only receive Level 1 container analysis.

Audio and video files are often large and personal. Analyzing them in your browser means they are never uploaded, and results appear faster.

Your browser reads the container (MP4, MOV, MP3, WAV, FLAC…) to extract codec, duration, bitrate and metadata, then decodes the audio and samples video frames to compute Level 2 indicators. Only an anonymous counter (content type and level) is sent to our server for statistics.