Quickstart

Run your first location analysis

Send one image to the Oceanir API and receive structured coordinates, an address, and the visual evidence behind the result.

REST APIJSON responseSet a 180s client timeout

Provision an API key

Eligible Pro accounts can create a personal key in Platform → API keys. Organization integrations need approved Enterprise access and organization-scoped credentials.

Request header

x-api-key: oceanir_4f9c2a7b1e08_1756...

Open API keys

Execute the analysis

POST an image URL. The depth parameter controls how much analysis the model performs.

POST/api/v1/geolocate

cURL

curl -X POST https://oceanir.ai/api/v1/geolocate \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "image_url": "https://example.com/photo.jpg",
    "depth": 2
  }'

depth 110s median · 24s p95

Classification

depth 26s median · 48s p95

Analysis

depth 383s median · 151s p95

Investigation

If a host blocks server-side image fetching, send a base64-encoded JPEG as image_b64. The maximum request payload is 20 MB.

Read the response

Use the returned coordinates and evidence in your own investigation workflow.

Illustrative D2 response · application/json

{
  "success": true,
  "analysis_id": "example-analysis-id",
  "location": {
    "latitude": 45.77,
    "longitude": 4.8459,
    "country": "France",
    "city": "Lyon",
    "neighborhood": "6th arrondissement",
    "address": "Lyon, France"
  },
  "confidence": 0.79,
  "reasoning": "Street signage and architecture suggest Lyon. Confirm the area with independent evidence.",
  "visual_clues": [
    {
      "label": "French street signage",
      "detail": ""
    },
    {
      "label": "Stone facades",
      "detail": ""
    }
  ],
  "model": "orca-1.5",
  "depth": 2,
  "timing_ms": 9120
}

analysis_idstring | nullSaved analysis identifier; may be null when history is not retained

locationobjectCoordinates, reverse-geocoded address, city, country, and ready-made map links

confidencefloatCertainty from 0 to 1 for the LEADING candidate only

reasoningstringVisual evidence used to reach the result

visual_cluesobject[]Objects with label and detail. Detail may be empty. Returned at depth 2 and above

forensic.candidatesarrayD3 only: competing candidates; optional and absent in this D2 example

forensic.contradictionsarrayD3 only: evidence against the leading candidate

calibrationobjectPublish policy for the run, including when to abstain. Depth 3 only

Try it live

Analyze an image without code

Authentication and key management