Machine-readable version of this same page: /llms.txt.
TerrAlert monitors user-defined geographic areas ("Areas of Interest") for land-use change, using Sentinel-2 optical satellite imagery and an AI change-detection model. Each area has an alert policy. Under "confidence", the default, a detection becomes an alert when its model score clears the area's minimum and the change is still there on the next clear image, compared with an earlier one. Under "corroborated", it becomes an alert only once an independent source that can see that kind of change has recorded it on or near the detection between its two images or a few days after (NASA FIRMS for fire, GLAD and RADD for forest loss, NASA flood maps for flooding); a signal of another kind nearby does not count. Construction and urban expansion have no such source at present. Detections raised by NASA FIRMS, GLAD or RADD themselves alert under either policy once their feed confidence reaches the area's minimum, labelled as that feed's report.
It watches for six categories of change: deforestation, construction and development, flooding, fire, vegetation loss, and urban expansion. Alerts are delivered by email or by webhook, with Slack and Microsoft Teams incoming webhooks detected automatically from the URL.
ESG and sustainability teams monitoring supply chains for deforestation risk; environmental NGOs and conservation organizations tracking protected areas; government agencies and regulators monitoring land-use compliance; insurance companies assessing wildfire, flood and land-change exposure; real estate and infrastructure developers tracking site activity.
A user draws a polygon on a map, or submits one as GeoJSON via the API. TerrAlert then fetches new Sentinel-2 imagery of that area as it becomes available, runs the change-detection model on each new clear pair of images, and once a finding passes the area's alert policy, sends an alert. Any size can be drawn: an area smaller than 10,000 m² or narrower than 100 m is watched through a box at least 100 m across, alerts for it report change at or next to it, and a change must be roughly 50 m across to be reported.
Detection speed is bounded by physics, not TerrAlert: Sentinel-2 revisits a given point roughly every 2 to 3 days, imagery takes about half a day to a day to become downloadable, and cloud cover removes many of the remaining passes. In practice, expect a detection within about a week of a real change, not same-day. Plan names such as "daily" and "realtime" describe how often TerrAlert checks for new imagery, not how often new imagery exists.
Flat monthly pricing, no per-scene or per-image fees. Prices are in US dollars and exclude tax, which is added at checkout where it applies. Every plan runs the full detection pipeline; plans differ in how many areas can be watched, how large each may be, how often it is scanned, and which alert channels are included.
Free, Starter, Growth and Pro are self-serve: create an account, choose a plan, and complete Stripe Checkout. No sales call or human approval is required for these four plans. Enterprise is a negotiated price, requested via the contact-sales form or the in-app Enterprise pricer, and paid the same way, a Stripe Checkout session, once a price is agreed.
Every purchase, on every plan, requires the signed-in account holder to complete Stripe's own checkout and confirm the charge. No integration, API call or automated agent can complete a purchase on a user's behalf without that explicit, authenticated confirmation. There is no endpoint that bypasses it.
API access, meaning programmatic use of the REST API with an API key, is included on Pro and Enterprise; the web app itself uses the same API on every plan. JSON in and out, versioned under /v1, scoped and revocable API keys created on the API keys page and sent as a bearer token, GeoJSON for area boundaries, and Cloud-Optimized GeoTIFF tiles for imagery. The written guide at /docs/api-guide lists the endpoints and rate limits, and the OpenAPI document is published at /api/openapi/v1.json.
Outbound webhooks, available on Starter and above (not on Free), are HMAC-SHA256 signed so a receiver can verify authenticity; see /docs/webhook-guide. Slack and Microsoft Teams incoming-webhook URLs are recognized automatically and receive a plain chat message instead of the signed JSON payload.
Connections are encrypted with TLS 1.2 or 1.3. Stored webhook signing secrets are encrypted at field level; encryption at rest beyond the hosting provider's disk defaults is not independently verified. Accounts support multi-factor authentication. Account holders can export their data and delete their account from the Account page. Data is stored in the United States. Full detail: /security, the privacy policy and /subprocessors.html.
Detection is not real-time in the sense of minutes: it is bounded by satellite revisit frequency and cloud cover, typically about a week end to end for optical detection. TerrAlert does not publish a single headline accuracy percentage, because a benchmark score on a research dataset does not represent real-world performance across every location. Production alerting instead relies on each area's alert policy: a public signal recorded nearby, or a user-configurable minimum model score with a persistence check. There is no contractual uptime guarantee unless an Enterprise customer signs a written SLA.
Support scales with plan: email on a best-effort basis on Free, email with a target first response of 2 business days on Starter and Growth, priority email with a target of 1 business day on Pro, and a named contact with response targets agreed in the order on Enterprise. These are targets, not guarantees.
This page reflects the production system as of its last deployment. See /changelog for what changed and when, and /llms.txt for the same facts as plain text.
Related: Pricing · FAQ · Developer documentation · Security · Changelog · llms.txt: the same facts as plain text