Oceanir
  • Pricing
SalesStart free

News Photo Verification for Fact-Checkers

Is this photo from where they say?

Misattributed photos travel faster than corrections. Upload the image, get coordinates and a confidence score, and build a verification trail you can publish.

No GPS or EXIF required/Free surface scan/Audit-grade PDF on Pro/How it works

For fact-checkers

  • →

    Coordinates, confidence, and visual cues returned on every scan.

  • →

    HowTo + FAQPage structured data for verifiable methodology.

  • →

    Citation template included for transparent reporting.

  • →

    Pro plan: AES-256-GCM encryption, 30-day auto-cleanup, audit-grade PDF export.

Case study: a misattributed conflict-zone photo

“The photo circulated as evidence of an attack in City A. Orca placed the scene 1,400 km away, in City B, where the same building had been damaged in a separate incident two years earlier.”

Documented verification, newsroom OSINT desk

The claim

Caption stated the photo showed damage from a recent strike in City A. The image was shared by multiple accounts and picked up by two outlets before any verification was attempted.

The verification

Orca returned coordinates matching City B. Visual cues — a specific minaret profile and a distinctive utility pole design — matched street-level imagery from City B dated two years prior. The photo was real; the attribution was not.

Five-step verification checklist

A repeatable process. Run it on every photo before publication.

  1. 1

    Establish provenance

    Trace the photo to its first appearance. Who published it first, when, and with what caption. Reverse image search gives you the chain of republication. The original publisher and timestamp are your anchor.

  2. 2

    Geolocate the scene

    Upload the photo to Oceanir. Orca returns coordinates, a confidence score, and the visual cues it used. Compare those coordinates to the location claimed in the caption or headline.

  3. 3

    Cross-check visual cues against the claim

    Review the cues Orca flagged: signage language, vehicle plates, utility infrastructure, vegetation, architectural vernacular. Each cue is a checkable fact. A single contradiction is a flag; two is a likely misattribution.

  4. 4

    Verify the time window

    Shadow angle and length encode time of day and season. If the shadow geometry contradicts the claimed date or time, the photo was taken at a different moment than reported. Sun elevation calculators confirm what Orca reads.

  5. 5

    Document and cite

    Record the Orca result with coordinates, confidence, and cues. Screenshot everything. Use the citation template below. A documented verification trail is what separates a fact-check from an opinion.

Citation template

Copy this into your fact-check. Disclose the confidence score and date. Transparency is the standard.

Source: Oceanir visual geolocation analysis
Date: [YYYY-MM-DD]
Image: [description or hash]
Coordinates: [lat, lng]
Confidence: [X]%
Visual cues cited: [list cues]
Analysis URL: [oceanir.ai result link, if saved]

Run the check

Upload the photo in question. Free surface scan, no signup. Coordinates and confidence returned in seconds.

No GPS or EXIF required/Free surface scan/Audit-grade PDF on Pro/How it works

Frequently asked questions

How does Oceanir help fact-checkers verify news photos?

Oceanir geolocates the image pixels and returns coordinates, a confidence score, and the visual cues that drove the prediction. Fact-checkers compare those coordinates to the location claimed in the caption. A mismatch is documented evidence of misattribution.

Can Oceanir verify photos from conflict zones?

Outdoor scenes with distinct infrastructure — roads, buildings, terrain, signage — resolve with useful confidence. Heavily compressed or low-resolution uploads common in conflict reporting reduce precision but still produce region-level estimates. Always pair the Orca result with traditional OSINT cross-checks.

What confidence threshold should I trust before publishing?

Treat anything below 50% as a lead, not a finding. 50-70% supports a qualified claim ('the photo is consistent with a location in X region'). Above 70% supports a stronger claim with the visual cues cited. Always disclose the confidence score in your reporting.

Does Oceanir preserve the images I upload?

Uploaded images are processed to return the result and are not used for training. Pro and Teams plans add 100m coordinate rounding, AES-256-GCM per-user encryption, and 30-day auto-cleanup. See the privacy documentation for the full architecture.

Can I cite Oceanir in a published fact-check?

Yes. Use the citation template on this page. Disclose the confidence score, the coordinates, and the date of the analysis. Citing the tool transparently is standard practice and strengthens the verification trail.

How is this different from reverse image search?

Reverse image search finds where else the photo has appeared. Oceanir finds where the photo was taken. The two methods are complementary: reverse search establishes provenance, geolocation establishes ground truth. A thorough fact-check uses both.

01platform
  • Pricing
  • Geoestimation
  • How It Works
02resources
  • Blog
  • Documentation
  • Contact
03company
  • About Us
  • Security
04legal
  • Terms
  • Privacy
  • Product Boundaries
  • SaaS Agreement
  • Cookies

Notes from the verification desk. What we're learning about reading places from pixels. Occasional, no noise.

Try it on one image

Upload a photo. Watch it come back as a place.

Run a free analysis →Talk to us →
Summarize with AI
oceanir
© 2026 Oceanir, Inc. All rights reserved.
PrivacyTerms