Live health for every Captapi platform, judged by the most recent responses — not a daily average. If something goes sideways, it shows up here first and clears the moment it is fixed.
Live status, judged by the most recent requests
Health reflects each platform’s most recent scrape responses: a platform only shows an issue while its latest requests are failing, and turns green again as soon as a request succeeds. Only server-side failures (HTTP 5xx) count — client errors such as invalid URLs or wrong query parameters never affect health. Speech-to-text extract misses on /youtube/audio-transcript are scored separately and do not turn the YouTube row yellow. Instagram Explore misses (trending-reels, reels-search, hashtag-search, transcript) are the same class — a geo scrape 502 does not turn Instagram yellow while permalink details still succeed. Facebook Marketplace and event-search timeouts are the same: they do not turn Facebook red while page-details still succeed. Last updated Oct 30, 2026, 2:20 AM UTC.
Measured
Computed from our request logs every 15 minutes. Last updated Oct 30, 2026, 2:18 AM UTC.
| Platform | 24h success | 24h median | 7d success | 7d median | 30d success | 30d median | 30d p95 |
|---|---|---|---|---|---|---|---|
| All platforms | 99.5% | 7.0 s | 98.7% | 3.6 s | 98.8% | 5.0 s | 30.1 s |
| Kick | — | — | — | — | 100.0% | — | — |
| Kwai | — | — | low traffic | — | low traffic | — | — |
| GitHub | low traffic | — | low traffic | — | low traffic | — | — |
| low traffic | — | 94.1% | 19.4 s | 92.9% | 9.3 s | 31.7 s | |
| Rumble | — | — | low traffic | — | low traffic | — | — |
| TikTok | 100.0% | 4.4 s | 98.0% | 1.2 s | 99.2% | 1.9 s | 17.2 s |
| Twitch | — | — | — | — | low traffic | — | — |
| Bluesky | — | — | — | — | low traffic | — | — |
| Linkbio | — | — | — | — | low traffic | — | — |
| Spotify | — | — | low traffic | — | low traffic | — | — |
| ThreadsDegraded | 90.7% | 16.6 s | 97.1% | 14.8 s | 99.2% | 15.5 s | 45.6 s |
| X / Twitter | 100.0% | 896 ms | 100.0% | 797 ms | 97.2% | 2.8 s | 7.8 s |
| YouTube | 100.0% | 2.8 s | 99.3% | 2.2 s | 99.6% | 2.6 s | 13.3 s |
| 96.2% | 4.6 s | 98.6% | 3.7 s | 96.9% | 4.3 s | 74.0 s | |
| low traffic | — | 100.0% | 5.4 s | 93.7% | 5.8 s | 46.0 s | |
| Linktree | low traffic | — | low traffic | — | low traffic | — | — |
| Snapchat | — | — | low traffic | — | 100.0% | 3.3 s | 6.3 s |
| 99.8% | 9.2 s | 99.7% | 6.9 s | 99.6% | 8.5 s | 27.5 s | |
| — | — | low traffic | — | 100.0% | 68.0 s | 110.6 s | |
| SoundCloud | — | — | — | — | low traffic | — | — |
| Video files | low traffic | — | 70.6% | 5.1 s | 81.5% | 5.0 s | 18.0 s |
| Google Maps | — | — | low traffic | — | low traffic | — | — |
| TikTok Shop | low traffic | — | 100.0% | 5.2 s | 99.3% | 11.1 s | 64.9 s |
| Truth Social | — | — | — | — | low traffic | — | — |
| Google Ad Library | — | — | — | — | low traffic | — | — |
| TikTok Ad Library | — | — | low traffic | — | 67.3% | 10.6 s | 80.4 s |
| Facebook Ad Library | low traffic | — | 100.0% | 24.4 s | 99.1% | 19.6 s | 48.1 s |
| LinkedIn Ad Library | — | — | — | — | 97.7% | 28.9 s | 78.9 s |
Source. Every billed API call writes one row to our request log with the HTTP status the caller received and how long our server took to answer. A scheduled job reads those rows every 15 minutes and stores one snapshot; this page and the success-rate line on our homepage and pricing page read that same snapshot.
Success rate = responses with status below 400 ÷ (responses below 400 + responses 500 and above). A 503 or any other 5xx counts as a failure. Client errors (4xx) are left out of the rate: they usually mean an invalid URL, a missing parameter or a private or deleted profile. Over the last 30 days, 404 “not found” responses were 4.0% of all responses and other client errors were 2.6%. Some 404s may be our miss rather than a deleted profile; we cannot tell them apart yet, so they are reported here instead of being counted either way.
Retries. When we retry upstream internally, the caller still sees one response and it counts once, by its final status. If you retry a failed call yourself, each attempt counts separately.
Latency is measured from the moment our server starts handling a request until the response is ready. It does not include network time between you and us. Only successful responses that were fetched live count; cached responses are excluded. Median is the 50th percentile, p95 the 95th (linear interpolation).
Low traffic. A platform needs at least 30 counted requests in a window before we show a rate or latency for it. Fewer than that is marked “low traffic”.
Degraded means the last-24-hour success rate is below 95%, or the live check above shows the most recent requests failing.
Internal test traffic excluded. Requests from our own test and monitoring accounts are not counted.
Not counted. Requests rejected before they reach a platform (invalid API key, rate limit, request validation, insufficient credits) are either not logged or are client errors, so they do not affect the rate. Requests lost while the server restarts are not logged either.
Not measured. We do not run an external uptime probe, so this page does not claim an uptime figure. Success rate is not uptime.
Excluded periods
Endpoint data excluded before a fix
/v1/instagram/profile-search before Oct 11, 2026, 10:15 AM UTC: Upstream failures were returned as empty results before this date.Methodology version 1.
History
1 total
ResolvedJul 4, 4:32 PM UTC
Transient upstream 504s on Reddit endpoints were mitigated: the fallback chain now retries the final scraper and maps leftover failures to clean 502s. All Reddit endpoints are operational.
InvestigatingJul 4, 3:00 PM UTC
Reddit endpoints are having issues. Working on a fix.
Email support@captapi.com and a human will look at it — usually within hours.