Geo Estimate
Where was this captured?
Generate ranked location candidates from the visible scene, with coordinates, confidence, and the clues behind the estimate.
Enterprise only
Oceanir Visual Intelligence Suite combines Geo Estimate, Street Match, and Property Verify in one controlled workflow for claims, disputes, investigations, and customer-facing products. The workspace is delivered through Oceanir Unit at unit.oceanir.ai or a named customer host such as parakeet.oceanir.ai.
Oceanir Visual Intelligence Suite
Each module answers a different investigative question. Together they move an analyst from an unknown image to a documented location decision without collapsing uncertainty into a single unexplained pin.
Where was this captured?
Generate ranked location candidates from the visible scene, with coordinates, confidence, and the clues behind the estimate.
Which street or address fits?
Compare the source against street-level reference imagery and rank the strongest address candidates by visual agreement.
Does the media support the location claim?
Test a claimed property or address and return a reviewable verdict with supporting evidence, contradictions, and alternatives.
Street-level geolocation is unsolved. On the standard public benchmarks, the best published research places roughly one image in five within a kilometre of its true location. Any tool claiming near-certainty on arbitrary photographs is either quoting a contaminated benchmark or not telling you when it is unsure.
That matters less than it sounds, because most workflows do not need certainty. They need to know how much weight a result can carry. An adjuster deciding whether to open an investigation, an editor deciding whether to publish, a moderator deciding whether to remove a listing — each of them can act on a qualified answer and none of them can act safely on an unqualified one.
So we report confidence as we assess it and do not inflate it when the evidence is thin. A low score with populated contradictions is the system declining to guess, which is the behaviour you want in a file that someone may later have to defend.
A self-serve key is internal use only. An enterprise licence grants the right to display Oceanir results inside your own product, under your own name, to people who are your customers rather than ours.
Forensic-depth analyses return the evidence supporting a location and the evidence contradicting it, alongside the alternatives that were considered and rejected. A reviewer can audit the conclusion instead of taking it on trust.
Where you process images belonging to your own customers, we sign as a sub-processor under their controller agreements. Images are retained for 30 days for debugging and audit, then deleted, and are never used to train shared models.
Inference runs across two independent providers with automatic failover between them, so a single provider incident does not take your workflow down.
The suite is available only through an Enterprise agreement. The licence fee covers the rights and the paperwork: customer-facing output, use of the Oceanir name in your product, and a data processing agreement written for your processing shape. Usage is billed against included volume, at any analysis depth including full forensic.
Individual Geo Estimate workflows remain available on the pricing page. Street Match, Property Verify, batch processing, production API access, evidence reports, and audit trails are scoped to Enterprise volume and use case, so terms are agreed on a call rather than listed here.
Bring a rough monthly image volume and whether the feature is customer-facing, and we can put real numbers to a licence on the first call rather than afterwards.