Sleep Tracking API
Sleep data from every wearable, in one schema
Sleep stages, duration, efficiency and timing, normalised across every ring, watch and mattress sensor.
7/12
providers connectable today
4
webhook event types
Canonical unit
seconds for durations, ISO-8601 timestamps for boundaries
What sleep looks like through one API
Sleep is the metric developers most often assume is standardised and most often discover is not. Every provider reports a duration, a set of stages and some efficiency figure, but almost nothing about how those numbers are produced is shared between them. Two devices on the same person on the same night routinely disagree by 30–50 minutes of total sleep time and by considerably more on individual stages.
WearLink normalises the parts that can honestly be normalised — session boundaries, total and per-stage durations in seconds, efficiency as a percentage, and the timestamps that anchor a session to a calendar date — and leaves the provider-specific scores alongside rather than pretending they are interchangeable. You get one object shape whether the night came from an Oura ring, a WHOOP band or an Eight Sleep mattress.
The hard part is not parsing sleep; it is deciding which night a session belongs to and knowing when a provider has silently revised it. Both are handled below the API surface, so your code sees a single settled record per session and a webhook when one changes.
Where providers disagree
Stage taxonomies do not line up
Some providers report four stages (awake, light, deep, REM), some collapse light and deep into a single "core" bucket, and some emit only asleep/awake. Summing a four-stage night and a two-stage night into the same chart produces a comparison that looks meaningful and is not. WearLink maps to the common set and marks which stages were actually measured versus inferred, so you can exclude devices that cannot answer the question you are asking.
Which day does a night belong to?
A session running 23:40 to 07:10 is attributed to the start date by some providers and the wake date by others. Get this wrong and every weekly aggregate is off by one for roughly half your users. WearLink stores both the session boundaries and an explicit attribution date so you can pick a convention and apply it uniformly.
Nights get revised after you fetch them
Oura recalculates as the ring syncs more data; Fitbit restates when a device backfills. A night you read at 08:00 can legitimately differ by 09:00. Polling once and caching produces silently stale data. WearLink refetches on the provider notification and emits a webhook with the corrected record.
Naps are inconsistently included
Some providers emit a separate short session for a daytime nap; others fold it into the main sleep record or drop it entirely. If your product computes "hours slept per day", the naps decision changes the answer. The normalised record flags session type so you can include or exclude deliberately.
Webhook events for sleep
Summary: sleep.created
Series streams:
series.sleeping_breathing_disturbances.createdseries.breathing_disturbance_index.createdseries.respiratory_rate.created
The full catalogue is in the event catalogue.
FAQ
Sleep — frequently asked questions
Which wearables can WearLink pull sleep data from?
Are sleep stages comparable between Oura and WHOOP?
How do I get notified when new sleep data arrives?
Can I get breathing and respiratory data with sleep?
Get sleep from every device your users own
Free for up to 3 connected users. No credit card. Your first API key is issued the moment you sign up.