ഈ ഡെവലപ്പർ ഡോക്യുമെന്റുകൾ നിലവിൽ ഇംഗ്ലീഷിൽ മാത്രമാണ്. കോഡും API സ്വഭാവവും എല്ലാ പ്രദേശങ്ങളിലും ഒരുപോലെയാണ്.

Pagination

List endpoints return results in pages using cursor-based pagination. Cursors are stable under inserts and deletes (unlike page numbers), so you never skip or double-read rows while paging through a moving dataset.

The scheme

Two query parameters:

Parameter	Type	Meaning
`limit`	integer	Max items to return in this page.
`cursor`	string	Opaque pointer to the next page. Omit it on the first request.

And a consistent response envelope:

json

{
  "items": [ ... ],
  "next_cursor": "1718971801.482"
}

Field	Meaning
`items`	The page of results (newest-first for time-ordered lists).
`next_cursor`	Pass this back as `cursor` to fetch the next page. `null` means you've reached the end.

Fetching the first page

bash

curl "https://api.ai92.ai/v1/account/usage/log?limit=50" \
  -H "Authorization: Bearer ai92_live_YOUR_KEY"

json

{
  "items": [
    {
      "request_id": "b2c1...",
      "ts": "2026-06-21T10:14:22Z",
      "method": "POST",
      "endpoint": "/v1/contacts/reveal",
      "service": "contacts",
      "status": 200,
      "credits": 3,
      "latency_ms": 142,
      "trace_id": "7f3c...",
      "backend_module": "module-H"
    }
  ],
  "next_cursor": "1718966062.117"
}

Fetching the next page

Pass next_cursor from the previous response as cursor:

bash

curl "https://api.ai92.ai/v1/account/usage/log?limit=50&cursor=1718966062.117" \
  -H "Authorization: Bearer ai92_live_YOUR_KEY"

Keep going until next_cursor is null.

Looping through every page

python

import httpx

API = "https://api.ai92.ai/v1"
headers = {"Authorization": "Bearer ai92_live_YOUR_KEY"}

def iter_usage(limit=100):
    cursor = None
    while True:
        params = {"limit": limit}
        if cursor:
            params["cursor"] = cursor
        page = httpx.get(f"{API}/account/usage/log", headers=headers, params=params).json()
        yield from page["items"]
        cursor = page.get("next_cursor")
        if not cursor:
            break

for row in iter_usage():
    print(row["ts"], row["endpoint"], row["credits"])

javascript

const API = "https://api.ai92.ai/v1";
const headers = { Authorization: "Bearer ai92_live_YOUR_KEY" };

async function* iterUsage(limit = 100) {
  let cursor = null;
  do {
    const url = new URL(`${API}/account/usage/log`);
    url.searchParams.set("limit", String(limit));
    if (cursor) url.searchParams.set("cursor", cursor);
    const page = await (await fetch(url, { headers })).json();
    yield* page.items;
    cursor = page.next_cursor;
  } while (cursor);
}

for await (const row of iterUsage()) {
  console.log(row.ts, row.endpoint, row.credits);
}

Notes

Treat cursor as opaque. It encodes position; don't parse, construct, or persist meaning into it. Just echo it back.
next_cursor: null is the only end signal. Don't infer "done" from a short page — request until the cursor is null.
Filters compose with pagination. Endpoints that accept filters (e.g. service, status_class on the usage log) keep them stable across pages — keep passing the same filters alongside the cursor.
Pick a limit that fits your throughput. Larger pages mean fewer round-trips; the activity log defaults to 50.