Configuration
Manage system integrations, team access, and platform health.
API Integrations
Gemini AI Model
Checking…Powers the automated risk assessment summarization.
GEMINI_API_KEY=…Google Earth Engine
Checking…Core engine for deforestation analysis.
EE_PROJECT=…Status is read live from GET /integrations, which introspects the backend's environment. Keys are set in .env (or the Cloud Run service config) and never sent to the browser — only whether they are present.
System Health
Earth Engine Tiers
Evidence stages the risk engine runs on every satellite assessment. Stages are cumulative — all signals feed each verdict — so this panel reports what ran; it doesn't select anything.
Tier 2 — Direct forest layers
EU JRC GFC2020 V3 · Hansen/UMD Global Forest Change · JRC Tropical Moist Forest
Primary evidence: post-2020 change fractions from these layers drive HIGH RISK and NEEDS REVIEW verdicts.
Tier 3 — AlphaEarth Foundations
Google Satellite Embedding V1 annual (10 m, 64-band)
Per-pixel embedding change between the Dec 31, 2020 cutoff and the latest year, reported as share of plot area.
An experimental shade-coffee classifier over the same embeddings ships with the engine but is not enabled in this deployment.
A stage that fails is reported under “Engine notes” on the results and evidence pack — it never silently drops out of the verdict.
Team Access
Sample dataManas Choudhary
Admin • manas@harshvs.com
Harshvardhan Singh
Admin • connect@harshvs.com