How close the account runs to its runaway-protection ceiling (max_rps in GET /v1/account): per time bucket, the requests the ceiling counted, the busiest second (peak_rps), the average rate and the requests refused with rate_limited. The ceiling is a safety limit that stops a runaway client, not a plan limit: plans have no rate limit. Account-wide (every key and the dashboard). Buckets are listed oldest first, cursor-paged. Data is complete to within about 5 minutes. Free.
Auth: an API key with the account:read scope, or the dashboard session.
Test mode: A test key gets an empty report: test requests are not metered.
Example
curl 'https://staging.socialtap.dev/v1/usage/rate' \
-H "x-api-key: $SOCIALTAP_API_KEY"| Name | Type | Description | Example |
|---|---|---|---|
since optional | stringmax 40 chars | Start, inclusive: RFC 3339 or a date (UTC), rounded down to the bucket. Default: one full page of buckets before until. | — |
until optional | stringmax 40 chars | End, exclusive: RFC 3339 or a date (UTC). Default: now. | — |
bucket_width optional | string | Bucket size. Default 10m. At most 144 ten-minute, 168 hourly or 31 daily buckets per page. One of 10m, 1h, 1d | — |
limit optional | integermin 1 max 168 | Buckets per page (default and maximum: 144 ten-minute, 168 hourly or 31 daily). | — |
cursor optional | stringmax 200 chars | page.next_cursor of the previous page. | — |
Start, inclusive: RFC 3339 or a date (UTC), rounded down to the bucket. Default: one full page of buckets before until.
End, exclusive: RFC 3339 or a date (UTC). Default: now.
bucket_widthstringBucket size. Default 10m. At most 144 ten-minute, 168 hourly or 31 daily buckets per page.
One of10m1h1d
Buckets per page (default and maximum: 144 ten-minute, 168 hourly or 31 daily).
page.next_cursor of the previous page.
Response
200
Request rate against the runaway-protection ceiling
| Field | Type | Description |
|---|---|---|
data | object[] | This page's items. |
data[].average_rps | number | requests divided by the bucket's length in seconds (for the bucket now under way, by its seconds so far), to three significant figures. |
data[].ending_at | string (date-time) | Bucket end, exclusive. |
data[].peak_rps | integer | The most requests counted in any one second of the bucket (a UTC clock second), refused ones included; 0 without requests. Compare it with max_rps (GET /v1/account). |
data[].rate_limited | integer | Of requests, those refused with rate_limited (free) because the account was above its ceiling. |
data[].recorded | boolean | false for a bucket that ends before the account's request rate was first recorded (its first request since the rate has been measured): no data, rather than no requests. Always false for a test key. |
data[].requests | integer | Requests the ceiling counted, refused ones included: endpoint requests (Prefer: respond-async ones too) and the gateway's own routes, from every key and the dashboard. Not counted: a job's items (the job paces them) and test-mode endpoint requests (they have a ceiling of their own). |
data[].starting_at | string (date-time) | Bucket start, inclusive (RFC 3339 UTC). |
meta | object | The shared route meta (see the API reference introduction). |
page | object | |
page.end_reason | string or null | Why there are no more pages; null while has_more is true. end: the list ended, or reached since/until. platform_limit: the platform stops serving further pages although the list goes on. max_pages: SocialTap's cap on pages in one walk. New values may be added. A short or empty page is not the end: only has_more false is. One of end, platform_limit, max_pages |
page.has_more | boolean | True when another page exists; then next_cursor is set. An empty page with has_more true is not the end. |
page.next_cursor | string or null | Pass as cursor, with the same other parameters, to get the next page; null at the end. |
page.next_cursor_expires_at | string (date-time) or null | When next_cursor stops working, RFC 3339 UTC; null at the end. Don't store cursors. |
dataobject[]This page's items.
ShowHide 7 child fields
average_rpsnumberrequests divided by the bucket's length in seconds (for the bucket now under way, by its seconds so far), to three significant figures.
ending_atstring date-timeBucket end, exclusive.
peak_rpsintegerThe most requests counted in any one second of the bucket (a UTC clock second), refused ones included; 0 without requests. Compare it with max_rps (GET /v1/account).
rate_limitedintegerOf requests, those refused with rate_limited (free) because the account was above its ceiling.
recordedbooleanfalse for a bucket that ends before the account's request rate was first recorded (its first request since the rate has been measured): no data, rather than no requests. Always false for a test key.
requestsintegerRequests the ceiling counted, refused ones included: endpoint requests (Prefer: respond-async ones too) and the gateway's own routes, from every key and the dashboard. Not counted: a job's items (the job paces them) and test-mode endpoint requests (they have a ceiling of their own).
starting_atstring date-timeBucket start, inclusive (RFC 3339 UTC).
metaobjectThe shared route meta (see the API reference introduction).
pageobjectShowHide 4 child fields
end_reasonstringnullableWhy there are no more pages; null while has_more is true. end: the list ended, or reached since/until. platform_limit: the platform stops serving further pages although the list goes on. max_pages: SocialTap's cap on pages in one walk. New values may be added. A short or empty page is not the end: only has_more false is.
One ofendplatform_limitmax_pages
has_morebooleanTrue when another page exists; then next_cursor is set. An empty page with has_more true is not the end.
next_cursorstringnullablePass as cursor, with the same other parameters, to get the next page; null at the end.
next_cursor_expires_atstring date-timenullableWhen next_cursor stops working, RFC 3339 UTC; null at the end. Don't store cursors.
400
400 Bad Request.
| Code | HTTP | Retry | Meaning | What to do |
|---|---|---|---|---|
invalid_input | 400 | No | The request parameters are missing or malformed, or the URL is not a supported link for this endpoint. | Fix the parameter named in param (the message says how); for a link of the wrong kind, call the endpoint the message names. |
invalid_cursor | 400 | No | The cursor is malformed, expired, belongs to another endpoint, or the other parameters differ from the request that issued it. | Repeat the original request without cursor to start the list again. |
invalid_inputNo retryThe request parameters are missing or malformed, or the URL is not a supported link for this endpoint.
Fix the parameter named in param (the message says how); for a link of the wrong kind, call the endpoint the message names.
invalid_cursorNo retryThe cursor is malformed, expired, belongs to another endpoint, or the other parameters differ from the request that issued it.
Repeat the original request without cursor to start the list again.
401
401 Unauthorized.
| Code | HTTP | Retry | Meaning | What to do |
|---|---|---|---|---|
unauthorized | 401 | No | The x-api-key header is missing, unknown or revoked (or, on the dashboard's routes, the session cookie is missing or expired). | Send a valid API key in the x-api-key header; without one, get a key at details.signup_url (AI agents: POST details.agent_signup_url). |
unauthorizedNo retryThe x-api-key header is missing, unknown or revoked (or, on the dashboard's routes, the session cookie is missing or expired).
Send a valid API key in the x-api-key header; without one, get a key at details.signup_url (AI agents: POST details.agent_signup_url).
403
403 Forbidden.
| Code | HTTP | Retry | Meaning | What to do |
|---|---|---|---|---|
forbidden | 403 | No | The API key is valid but may not make this request: its account is suspended, its scopes do not include the route, it is waiting for its owner's email confirmation, it is past its expiry, or the request came from an IP its allowlist does not include; or the request used a rotated key's previous secret (or one no longer current) for something that needs the admin scope. A running job stops the same way when its account is suspended or the key that created it is revoked or expires. It is about the key, never the item (a private item is private_or_unavailable, also 403): branch on the code. | Do what details.reason says: if suspended, contact support; otherwise use a key whose scopes cover the route (details.required_scope) or the key's new secret after a rotation, have the owner confirm the emailed link, create a new key, or call from an allowed IP. |
forbiddenNo retryThe API key is valid but may not make this request: its account is suspended, its scopes do not include the route, it is waiting for its owner's email confirmation, it is past its expiry, or the request came from an IP its allowlist does not include; or the request used a rotated key's previous secret (or one no longer current) for something that needs the admin scope. A running job stops the same way when its account is suspended or the key that created it is revoked or expires. It is about the key, never the item (a private item is private_or_unavailable, also 403): branch on the code.
Do what details.reason says: if suspended, contact support; otherwise use a key whose scopes cover the route (details.required_scope) or the key's new secret after a rotation, have the owner confirm the emailed link, create a new key, or call from an allowed IP.
429
429 Too Many Requests.
| Code | HTTP | Retry | Meaning | What to do |
|---|---|---|---|---|
rate_limited | 429 | Yes, later | The account exceeded its runaway-protection ceiling: plans do not limit request rate, but a high per-account safety ceiling stops broken client loops. Support raises it on request. Without an API key (signup, the MCP server's tool listing and find_endpoint), requests are limited per IP address instead. | Wait for the number of seconds in the retry-after header, then retry; ask support to raise the ceiling if your sustained traffic needs more (calls without an API key: use one). |
rate_limitedYes, laterThe account exceeded its runaway-protection ceiling: plans do not limit request rate, but a high per-account safety ceiling stops broken client loops. Support raises it on request. Without an API key (signup, the MCP server's tool listing and find_endpoint), requests are limited per IP address instead.
Wait for the number of seconds in the retry-after header, then retry; ask support to raise the ceiling if your sustained traffic needs more (calls without an API key: use one).
A test key gets an empty report: test requests are not metered.
operationId usage.rate.list