What can we reasonably infer about this water?
Upload a photograph of a water source and its location. WaterLens examines the visible evidence, maps the surrounding area, researches public records, and returns an assessment that separates what is observed from what is documented and what is only inferred.
A photograph cannot measure chemistry or bacteria. WaterLens will say so — insufficient data is a valid result.
Three layers, never merged
Most tools collapse guesswork into a single verdict. WaterLens keeps the layers apart so you can judge the reasoning yourself.
- Observation
“The water appears brown and cloudy.”
Evidence directly visible in your photograph.
- Evidence
“A government report documented pollution concerns in this watershed.”
Information retrieved from external sources, with links.
- Inference
“These findings raise concern but do not prove contamination at your exact location.”
A conclusion drawn from observations and evidence together.
How an investigation runs
Four stages, streamed to your screen as they happen. No fake progress bars.
- 01
Visual analysis
A vision model reports only what is visible: colour, clarity, turbidity, particles, foam, algae, debris.
- 02
Geographic context
OpenStreetMap data locates nearby waterways, industry, agriculture, landfills and treatment facilities.
- 03
Web research
A research model plans queries, then reads real search results and keeps every source URL.
- 04
Reasoning
A final model synthesises a structured assessment, with confidence and stated limitations.
Every claim keeps its source
Assessments link back to the pages they came from, so you can verify the evidence rather than trust the model.