BuildonConstellation
Everything engineers need to integrate — contracts, auth, and deployment details — without the positioning. Same API in commercial cloud and AWS GovCloud (US). If you want the why, read the platform pages; this page is the how.
Try it
The API, before the sales call
curl "https://api.constellation.space/predictions\
?model_family=snr&live=true\
&as_of_utc=2026-07-09T17:21:00Z\
&link_ids=sat-14__gw-01" \
-H "x-api-key: $CONSTELLATION_API_KEY"Response
// Send a request to see the response contract.
// p10/p50/p90 quantiles + provenance on every prediction.API surface
POST/telemetry
Ingest normalized fleet telemetry. Synchronous write — 202 means durable.
GET/predictions
Forecasts by model family (e.g. snr, gs_select). Batched link ids, conformal quantiles.
GET/topology
Query fleet topology and link state as-of a timestamp, with freshness bounds.
POST/simulations
Run constellation and contact simulations against deterministic physics.
POST/optimize
Contact scheduling, station selection, and routing as callable services.
First calls
# Ingest a telemetry point
curl -X POST https://api.constellation.space/telemetry \
-H "x-api-key: $CONSTELLATION_API_KEY" \
-H "content-type: application/json" \
-d '{"records":[{"measurement":"link",
"time":"2026-07-09T17:20:00Z",
"tags":{"entity_id":"sat-14__gw-01"},
"fields":{"snr_db":12.4}}]}'
# Forecast link quality (conformal p10/p50/p90)
curl "https://api.constellation.space/predictions\
?model_family=snr&live=true\
&as_of_utc=2026-07-09T17:21:00Z\
&link_ids=sat-14__gw-01" \
-H "x-api-key: $CONSTELLATION_API_KEY"Link ids use the source__target convention. Prediction responses include per-horizon quantiles; treat p10–p90 as the planning envelope, not p50 as truth.
Engineering guides
Learn from the engineers who built Constellation
All technical articles & deep dives →Developer tooling
Meet your stack where it already is
Integration guides
Authentication
Per-tenant opaque API keys, verified against server-side argon2 hashes — keys are never derivable and never leave your tenant. Scoped per capability (telemetry:write, predictions:run) and minted from Console settings.
Telemetry ingestion
One write contract for every vendor: measurement, timestamp, tags, fields. Writes are synchronous and durable on acknowledgment; ingested points are immediately queryable.
Predictions
Request forecasts for batches of links at a given as-of time. Responses carry p10 / p50 / p90 conformal quantiles per horizon so your systems can schedule against bounded uncertainty.
Rate limits & batching
Per-tenant and per-IP limits are enforced at the gateway. Prediction requests accept batched entity ids — one call per tick, not one call per link.
Environments
Dedicated development, staging, and production environments with promotion gates. Enterprise and government tenants deploy to isolated stacks — see deployment modes.
Errors & auditability
Structured error envelopes with stable codes. Every authenticated call is attributable to a key and tenant, and lands in the audit stream.
Deeper architecture — tenant isolation, VPC design, data boundaries, deployment topologies — lives in the technical whitepaper and on Standards & Compliance.