Free sandbox
Test requests and sample responses without a payment card. It does not analyse your files.
AI media checks for developers
Deepfake detection API for images and video.
Check uploaded images and short videos for AI-generation signals. Retrieve results asynchronously as JSON, text or a PDF with your logo, then send uncertain files for human review. Start with fixed sandbox examples before connecting live processing.
Basic €0.06 / Full €0.12 - EUR per completed image · Video pricing and the 60-second limit
Explore a complete sample report
REPORT V1 · EXAMPLE JSON
{
"schemaVersion": "deepfakepolicy.report.v1",
"reportId": "DFP-SAMPLE-IMAGE-0001",
"generatedAt": "2026-09-13T12:00:00.000Z",
"source": {
"modality": "image",
"sha256": "55c824c6118dfb18ef5156c3e7c6b5bf4cf9cd53ef44c3f1b5f5efa27c45910e",
"name": "demo-portrait.jpg",
"mediaType": "image/jpeg",
"format": "JPG",
"url": null,
"sizeBytes": 63578,
"pageCount": null,
"characterCount": null
},
"analysis": {
"mode": "quick",
"status": "complete",
"creditsUsed": 1,
"method": "DeepfakePolicy evidence synthesis"
},
"evidencePath": [
{
"id": "provenance",
"label": "Provenance",
"status": "limited",
"summary": "No verified origin credential or complete chain of custody was available for this report.",
"signals": []
},
{
"id": "metadata",
"label": "Metadata",
"status": "reviewed",
"summary": "Available format, coverage and metadata observations were reviewed.",
"signals": [
{
"label": "Submitted format",
"finding": "JPG",
"score": null,
"scoreLabel": null
}
]
},
{
"id": "media-signals",
"label": "Media signals",
"status": "not-available",
"summary": "No separate human-readable media observation was available for this modality.",
"signals": []
},
{
"id": "model-signals",
"label": "Model signals",
"status": "reviewed",
"summary": "The returned classifications are listed separately, including unavailable checks and any segment findings.",
"signals": [
{
"label": "AI-generated image",
"finding": "Indicators found",
"score": 82,
"scoreLabel": "82% signal score"
},
{
"label": "Deepfake face manipulation",
"finding": "No strong signal in this example",
"score": 6,
"scoreLabel": "6% signal score"
}
]
},
{
"id": "web-evidence",
"label": "Public web evidence",
"status": "not-available",
"summary": "Automated public-source research was not included in this analysis.",
"signals": []
}
],
"verdict": {
"code": "suspicious",
"label": "Suspicious signals detected",
"aiLikelihood": 82,
"confidenceBand": "not-rated",
"summary": "The analysed media contains signals that require review. The individual findings identify what to examine.",
"rationale": [
"The AI-generation detector returned an elevated signal.",
"No independently verified capture history is supplied in this example."
]
},
"recommendedActions": [
"Preserve the earliest available, highest-quality original and its surrounding context.",
"Confirm the publisher, identities and important claims through an independent channel.",
"Compare key frames, audio segments or images with an attributable original source.",
"Escalate high-stakes use to a qualified forensic reviewer before acting."
],
"limitations": [
"This is a probabilistic assessment, not proof of authenticity, authorship or intent.",
"Compression, editing, transcription, translation, short samples and unseen generation methods can change results.",
"A low score does not certify authenticity, and a high score does not identify a creator.",
"Public-source research can miss private, deleted, paywalled or unindexed material."
],
"privacy": {
"originalStoredWithReport": false,
"researchContributionSeparate": true,
"reportVisibility": "response-only",
"statement": "The report contains normalized findings, not the submitted source. Any optional research contribution is governed and deleted separately."
}
}This fictional image example illustrates a standard check. Source research is a separate option; available evidence depends on the file and selected analysis mode. aiLikelihood is a model signal, not the probability of fraud or a guarantee of authenticity. The API field names stay the same across integrations.
Deepfake detection API pricing and scope
Image detection API
Basic: €0.06 per completed image. Full: €0.12, including a separate visual analysis. Both provide the recorded findings for review.
API prices, currencies and options →Video detection API
€0.04 per started second, with a €0.50 minimum per completed video. Submit supported clips up to 60 seconds, one job per file.
Request formats and upload limits →Test the integration before live analysis
Free sandbox fixtures test requests and reports. For your own files, activate company billing and set a spending limit. Completed inconclusive reports are billable; processing failures are not.
Start with the sandbox quickstart →Example: 1,000 completed images: Basic €60.00 · Full €120.00. Usage example, not a minimum commitment or free allowance. Excludes source research and applicable taxes.
Prices shown are for EUR billing. Source research is an optional, separately priced service. JSON, text and PDF downloads use the same saved analysis; exporting it does not run or charge for another check.
Evaluate the files your workflow actually receives
Freeze your labelled file list before scoring. The offline utility counts unanswered and failed checks and produces a draft; it does not run the detector or establish independent accuracy.
A result your team can act on.
Add visual analysis with Full.
Full image results include a separate automated visual review alongside specialist detector signals. The report explains supporting findings, disagreements, limitations and what to review next without folding the visual model’s self-rating into the detector score.
Fit your existing workflow
Use asynchronous jobs, versioned reports and signed webhooks. Keep your upload flow and review queue in your own product. JSON, text and PDF share the same recorded findings; exporting another format does not trigger another analysis charge.
Start small, control usage
Test the integration with free sample responses. Add a payment card before checking real files, set a spending limit and track usage in the API console.
Evaluate the API on the files your team receives.
A successful request tests the integration. A useful pilot also tests false alerts, missed findings and the reviewer’s next action.
Keep a labelled set
Document the source of genuine and generated files before you see the results. Separate original files from resized, recompressed and screenshot copies.
Record every outcome
Keep the input hash, analysis date, available findings and coverage. Count failures and inconclusive reports separately; an export is the same recorded check.
Choose a review rule
Measure useful escalations and unnecessary review against your own labels. Keep the final decision with your team, and set a spending limit before live processing.
Choose the right first check.
For one file, use the image or video checker and read the report before writing integration code. For a recurring upload flow, use the API: test requests with fixed sandbox examples, then send your own files with a Live key after setting up billing.
Choosing a provider? Use the deepfake detection API evaluation checklist to plan a representative test set, review false alerts and compare integration requirements.
Report formats
One report. Your presentation.
Download a ready-made PDF with your company’s branding. JSON and TXT provide the same recorded findings for your own interface and document templates.
JSON
Structured results for your application, with a versioned format.
Text export
Plain-text results for CRM notes, case records, emails or document templates.
PDF with your logo
Save your company name, logo and accent colour in Settings. PDFs downloaded through the API automatically use your workspace’s branding.
How it works
From upload to report.
Create a job
POST /v1/jobs returns a job ID and a protected upload path.
POST /v1/jobsUpload your file
Upload a supported image or a video up to 60 seconds from your backend.
PUT /v1/jobs/{id}/uploadProcess asynchronously
Keep your application responsive while the check runs. Retrieve status by polling or wait for a signed webhook.
GET /v1/jobs/{id}Use the report
Download the PDF with your workspace’s saved branding. Use result.evidenceReport, JSON or TXT to show the same findings in your own interface.
JSON · TEXT · PDFOne API. Different reasons to check.
Connect the same photo and video checks to the point where your business receives media. Your team decides what happens next.
AI image detection for insurance claim photos
Check claim photos and short videos for AI signals, then keep media findings separate from photo-document consistency and your claims decision.
See the workflow
Profile photos & communities
Check portraits for signs of AI generation when users add or update a profile photo.
See the workflow
AI image detection for marketplace listings
Check listing photos and short clips for AI signals. Review selected files in a company batch or connect the API to your existing upload and moderation flow.
See the workflow
Images within documents
Check photos extracted from documents for signs of AI generation.
See the workflowAPI v1
Controls for your integration.
Scoped keys
Create, expire and revoke company keys. Separate sandbox and live access. Key secrets are shown once.
Signed webhooks
Receive HMAC-signed notifications with retries, or poll for results. Use a versioned report schema.
Spending controls
Pay after use with no monthly subscription. Set a monthly spending limit; checks in progress count towards it. Failed checks are not billed.
Retention controls
Choose one hour to seven days, with a 24-hour default. Request deletion through the API and inspect its status.
Know what was checked.
The report explains detection signals in the submitted file. Combine it with the source and other case evidence; your team decides the next step.
Test the integration, then check your files.
Use the free sandbox to verify requests, fixed sample results and webhook handling. It does not evaluate your files. Use an individual check or a live API key to assess results on your own media. A completed inconclusive report is a billable check; a processing failure is not.
OpenAPI 3.1 · JSON Schema · Python / JavaScript / curl · Reports with your logo
Choose the API for your workflow
AI image, video & deepfake detection
Run image and video checks, track each job and retrieve its detector report.
POST /v1/jobsView integration guide
Photo & document Cross-check
Compare supplied photos, PDFs, documents and tables for consistent identifiers, amounts and dates. Keep source references with each finding.
POST /v1/comparisonsExplore Cross-checkConnect your workflow
Cross-check from the system your team already uses.
Import an n8n workflow to send a case, follow its progress and return the findings with your original case ID.
Retry delivery without another check
The resume action reads the saved job. It does not submit the files for analysis again.
n8n workflow
Send a case and receive the findings.
- Send photographs and documents from a new case.
- Receive the job ID, then follow the check automatically.
- Map the result back to the original case in your system.
Up to 6 files · 10 MiB per case · limited automatic retries
You configure the destination and its credentials. No CRM is connected automatically. Completed live Cross-checks are billable; the sandbox does not evaluate these files.
Questions before integrating detection
What does the image and video detection API return?
A completed job provides a structured evidence report with the detector findings, available file information, interpretation, limitations and suggested next steps. JSON, text and PDF exports support different review workflows.
Can I process a whole folder through the API?
Submit a job for each supported file and track its status or webhook. Company concurrency and spending limits apply across jobs. You can also select files in the company browser workspace without building an integration. Browser batches accept photos and short MP4, MOV or M4V videos, with a limit of 20 MB and 60 seconds per video.
Does the API detect every Photoshop edit or certify identity?
No. It assesses detectable signals associated with AI generation and manipulation. Ordinary retouching may not produce those signals, and the result is not a liveness or identity check.
How do I test before processing customer files?
Create a business workspace and key, then use fixed sandbox examples to test the request, upload and reporting flow. Connect company billing for real-file analysis and evaluate detection separately on representative media. No sales call is required.
From first test to your own integration.
See a report
Try your own file with a personal account, or explore the sample report before signing up.
Connect your application
Create a company workspace, issue a free sandbox key and test requests, sample results and webhooks.
SANDBOX · POST /v1/jobsEnable real-file checks
Add a payment card, set your monthly spending limit and use your Live key to check photos or videos. Keys are available before billing is set up.