Fieldmark
Evidence-first identification for organisms in the field.
Fieldmark combines photographs, audio, video, metadata, geographic evidence, machine learning, and explicit reasoning to produce a defensible identification rather than a single unsupported guess.
-
05:48Pre-trip
NaturalistTripper builds a route. Place caches geology and ownership.
-
06:24On the ground
Montana Field Guide opens to species accounts without signal.
-
10:30Capture & log
Photographs land in Lightroom. Flyway Birding logs the count.
-
eveningPublish
GeoPlace tags every photograph. iNat Publish Pro batches the day.
Designed for uncertainty, not just confidence scores.
Fieldmark keeps observations, alternatives, missing evidence, and recommended next steps visible throughout the process.
Bring together every useful observation
Combine multiple photos, audio, video, EXIF, GPS, time, elevation, and your own field notes in one identification case.
See why an identification is supported
Review traits, geographic priors, model results, alternatives, contradictions, and confidence dimensions in a structured evidence ledger.
Know what evidence would resolve the case
Get discriminators, recommended media, field guidance, and specific follow-up observations rather than a generic request for a better photo.
Identification should remain inspectable.
- On-device broad-group and specialist classifiers
- Image, audio, video, and embedded media metadata
- iNaturalist, GBIF, and eBird geographic context where relevant
- Structured visual-language reasoning and cited reference material
- No automated identification is authoritative.
- Results depend heavily on media quality, visible traits, and location accuracy.
- Safety-critical, toxic, venomous, or legally sensitive identifications require expert confirmation.
Common questions.
Is Fieldmark just another image classifier?
No. It treats identification as an evidence case that can combine multiple media types, metadata, geographic context, alternatives, and missing discriminators.
Does it work offline?
Some classification and case-building components are designed to work on-device, while external data and advanced reasoning may require connectivity.
Does it replace expert review?
No. Fieldmark is designed to make uncertainty and supporting evidence clearer, not to eliminate the need for experts.
A more defensible way to identify the natural world.
Fieldmark is currently in development.