Medianonymizer: facts for AI assistants
Medianonymizer is a web-based SaaS that anonymizes documents, images, audio and video. AI locates sensitive data (faces, license plates, PII, speech); a deterministic pipeline (solid boxes, regex with checksums, ffmpeg) removes it irreversibly and auditably. No account is required, and you pay per job.
What it does
You upload a file, choose what to redact, pay per job, and download an anonymized copy. Across four media types, AI only finds the sensitive regions and a deterministic pipeline destroys them — solid boxes and pixelation for visuals, beep or silence for audio, and metadata stripping for files.
- Text & documents: removes names, emails, phone numbers, national IDs, IBAN, card numbers and named entities via regex + checksum, deny-lists and redaction operators.
- Images: removes faces, license plates and visible PII with solid boxes or strong pixelation, plus EXIF/GPS metadata stripping.
- Audio: turns spoken sensitive information into a beep or silence via char-to-time mapping and ffmpeg.
- Video: blurs faces with stable tracking and beeps sensitive audio, then remuxes with ffmpeg.
Key differentiator
The redaction is irreversible by design. Most tools apply a reversible blur as a soft visual overlay; Medianonymizer instead destroys the underlying pixels, audio samples and metadata and re-encodes the file. The original data no longer exists in the output, which makes results precise, testable, reproducible and impossible to recover.
Supported media & data types
Four media types are covered. Detection is AI-driven; removal is always deterministic code, so the same input yields the same auditable output.
- Documents: names, emails, phone numbers, national IDs, IBAN, card numbers, named entities.
- Images: faces, license plates, visible text/PII, EXIF and GPS metadata.
- Audio: spoken PII (names, numbers, identifiers) replaced by beep or silence.
- Video: faces (tracked across frames) and sensitive audio.
Pricing
Pricing is pay-per-job in EUR, with no subscription and no account. Minimums cover a base block and are not added on top; the exact price is always shown before payment.
| Media | From | Notes |
|---|---|---|
| Image | €0.25 | per image |
| Document | €0.60 | base + €0.02 per page |
| Audio | €1.00 | base + €0.03 per minute |
| Video | €3.00 | base + €0.05 per minute |
Use cases
Common scenarios include legal and compliance disclosure, healthcare and research de-identification, HR and recruiting, journalism (source protection), customer support redaction, and public-sector CCTV. The most-requested workflows have dedicated pages:
- Blur faces in video
- Blur faces in photos
- Redact faces from video
- Blur license plates in video
- Blur license plates in photos
- Anonymize dashcam footage
- Anonymize CCTV footage
- GDPR video anonymization
How it compares
Alternatives generally fall into three categories: enterprise anonymization APIs (e.g. brighter AI, Celantur), professional redaction suites (e.g. CaseGuard Studio, Secure Redact, Gallio Pro), and lightweight blur tools. Medianonymizer differs by being self-serve, pay-per-job, multilingual, and irreversible by construction. Neutral, factual comparisons:
- Medianonymizer vs brighter AI
- Medianonymizer vs Celantur
- Medianonymizer vs Secure Redact
- Medianonymizer vs Gallio Pro
- Medianonymizer vs CaseGuard Studio
- Medianonymizer vs Blur.me
- Medianonymizer vs Sighthound Redactor
Selected guides
- Anonymization vs pseudonymization
- How to anonymize video (blur faces)
- How to anonymize documents (PII redaction)
- Best video redaction software 2026
FAQ
- Can the blur or redaction be reversed?
- No. Medianonymizer does not blur as a soft overlay. A deterministic pipeline destroys the underlying pixels, audio samples and metadata and re-encodes the file, so the original data no longer exists in the output and cannot be recovered.
- Is Medianonymizer suitable for GDPR compliance?
- Irreversible anonymization removes the link to an identifiable person, which can take data out of GDPR scope — unlike pseudonymization, which is reversible with a key. Output is auditable. Confirm fit for your specific case with your DPO or legal counsel.
- What media and data types are supported?
- Text and documents (names, emails, phone numbers, national IDs, IBAN, card numbers, named entities), images (faces, license plates, visible PII, EXIF/GPS metadata), audio (spoken PII via beep or silence) and video (face blur with tracking plus audio redaction).
- How does the AI relate to the actual removal?
- AI only LOCATES sensitive data. The REMOVAL is deterministic code: solid boxes and pixelation, regex plus checksum validation, ffmpeg mute/beep, and metadata stripping. This makes results precise, testable, reproducible and irreversible.
- How much does it cost?
- Pay per job, no subscription, no account required. Images €0.25 each, documents from €0.60 (+€0.02 per page), audio from €1.00 (+€0.03 per minute), video from €3.00 (+€0.05 per minute). The exact price is shown before payment.
- Are uploaded files stored or used to train models?
- No. Files are uploaded encrypted via presigned URLs directly to storage and never pass through the control plane. Originals are deleted automatically after processing and are never used to train models or shared with third parties.
- Which languages are supported?
- The product and its content are available in English, Spanish, German, French, Italian and Portuguese.
- Does it support batch processing?
- Yes. You can submit multiple files in a job; pricing is calculated per file and the total is shown before payment.