1

Select Your File

Choose any audio or video file from your device. We accept MP3, MP4, WAV, MOV, AVI, MKV, FLAC, OGG, AAC, WMA, AMR, MPEG, WMV, FLV, WEBM, M4A, 3GP, and more. Up to 500 MB per file.

2

Browser-Side SHA-256 Hash

The moment you select a file, your browser computes a SHA-256 cryptographic hash using the Web Crypto API. This hash is your file's unique fingerprint — computed entirely on your device before any data is uploaded. It's displayed on screen so you can record it independently.

3

Secure Upload & Payment

Your file and its hash are uploaded together over encrypted channels. You'll be taken to a secure Stripe checkout — per-minute pricing with no accounts, no subscriptions. The hash travels with your file through every stage of processing on our air-gapped infrastructure.

4

Download Your Verified Package

Once processing completes, you receive your full deliverable package: a speaker-labeled transcript with confidence scores, segment data with word-level timestamps, a PDF document with embedded metadata, and your SHA-256 hash verification file. You also receive a unique retrieval code to access your encrypted archive for one year.

Privacy & Data Handling

No Accounts

No signup, no login, no profile. Ever.

Source Files Deleted

Your audio/video is permanently deleted the moment transcription completes.

Output Archived 1 Year

Deliverables are hashed, encrypted, and archived for a minimum of one year.

No Follow-Up Emails

We don't store your email. No marketing, no newsletters, no spam.

Sample Output Preview

PERCEPTIVE VISION — TRANSCRIPTION OUTPUT
Job ID: 924355b4-xxxx-xxxx-xxxx-xxxxxxxxxxxx
File: bodycam_incident_47.mp4
Duration: 5 min 13 sec  |  Words: 367  |  Speakers: 3  |  Processed: 2026-03-20
[Speaker 1 — Officer]
[00:00:01] 97% Central, I'm initiating a traffic stop on a white sedan...

[Speaker 2 — Dispatch]
[00:00:14] 94% Copy, unit 47. Proceed with caution.

[Speaker 1 — Officer]
[00:00:22] 91% Driver, license and registration please.

[Speaker 3 — Driver]
[00:00:31] 38% — NOT LEGIBLE [inaudible — background noise]

[Speaker 1 — Officer]
[00:00:45] 88% Do you know why I pulled you over?
SHA-256: a3f8c9e2b1d4f5a6e7c8d9b0a1f2e3d4c5b6a7...matches uploaded file
This transcript was produced by automated speech processing. Confidence scores are estimates. All output should be verified by a qualified reviewer before use in any legal, regulatory, or official context.

What You Receive

Every transcription produces a complete deliverable package. All files include your SHA-256 hash and processing metadata.

PDF Document

Professional formatted output with file metadata, speaker labels, confidence scores, timestamps, hash on every page, and printed disclaimers.

Speaker-Labeled Transcript (.txt)

Full timestamped transcript with speaker separation, confidence percentages, and "not legible" markers for uncertain segments.

Segment Data (.json)

Word-level timing data with speaker IDs, confidence scores per word, and segment boundaries. Built for programmatic analysis.

SHA-256 Hash File

Standalone verification file containing your source file hash, output file hashes, and processing timestamps.

Retrieval Code

Your unique code to access the encrypted archive of your deliverables for one year. Store it securely — lost codes cannot be recovered.

Key Events Timeline

Automatically flagged notable moments with timestamps and speaker attribution. Includes low-confidence and inaudible segment alerts.

Supported Formats

MP3MP4WAV MOVAVIMKV WEBMFLACOGG AACM4AMPEG WMVFLVMPG WMAAMR3GP M4VMPEG4

What's Coming Next

Perceptive Vision is expanding beyond transcription into a full forensic analysis platform.

Video Forensics

Frame-by-frame analysis, metadata extraction, splice and tampering detection.

Audio Forensics

Voice isolation, spectral analysis, edit point detection, noise profiling.

Document Forensics

PDF integrity verification, metadata analysis, alteration and redaction detection.

Sealing Engine

Cryptographically sealed containers for tamper-proof evidence packaging and chain-of-custody.

Ready to try it?

Upload Your File