Meta Information

Every API response includes HTTP headers with rate limit data, request IDs, and token limits. Access them with the meta() method on any response.

use Anthropic\Laravel\Facades\Anthropic;

$response = Anthropic::messages()->create([
    'model' => 'claude-sonnet-4-6',
    'max_tokens' => 1024,
    'messages' => [
        ['role' => 'user', 'content' => 'Hello, world'],
    ],
]);

$meta = $response->meta();

$meta->requestId; // 'req_012nTzj6kLoP8vZ1SGANvcgR'

Include this when contacting Anthropic support. A common Laravel pattern is to log it alongside every successful response:

Log::info('Claude response', [
    'request_id' => $response->meta()->requestId,
    'input_tokens' => $response->usage->inputTokens,
    'output_tokens' => $response->usage->outputTokens,
]);

$meta->requestLimit->limit;     // 3000
$meta->requestLimit->remaining; // 2999
$meta->requestLimit->reset;     // '2026-04-14T13:29:17Z'

$meta->tokenLimit->limit;     // 250000
$meta->tokenLimit->remaining; // 249984

$meta->inputTokenLimit->limit;
$meta->inputTokenLimit->remaining;

$meta->outputTokenLimit->limit;
$meta->outputTokenLimit->remaining;

If your organization is on Priority Tier, two extra rate-limit buckets appear as typed properties. They stay null when the request didn't draw from Priority capacity:

$meta->priorityInputTokenLimit?->limit;
$meta->priorityInputTokenLimit?->remaining;
$meta->priorityInputTokenLimit?->reset;

$meta->priorityOutputTokenLimit?->limit;
$meta->priorityOutputTokenLimit?->remaining;
$meta->priorityOutputTokenLimit?->reset;

Useful for monitoring whether a request actually hit your Priority allocation or spilled into the shared standard pool. A quick Laravel-flavored check:

$meta = $response->meta();

if ($meta->priorityInputTokenLimit === null) {
    Log::warning('Claude request did not use Priority Tier', [
        'request_id' => $meta->requestId,
    ]);
}

If you're doing high-volume work, check the remaining requests and throttle before you hit a 429:

$response = Anthropic::messages()->create([...]);

$meta = $response->meta();

if ($meta->requestLimit->remaining < 10) {
    $reset = Carbon::parse($meta->requestLimit->reset);
    Cache::put('anthropic.throttle', true, $reset);
}

Your next job can check the cache flag before making a request.

Call meta() on the stream itself:

$stream = Anthropic::messages()->createStreamed([...]);

$meta = $stream->meta();

The meta information comes from HTTP response headers, which arrive at the start of the stream.

Same pattern works on batch results:

$results = Anthropic::batches()->results('msgbatch_...');

foreach ($results as $individual) {
    // Process results
}

$results->meta();

For the full header list and the raw toArray() format, see the Meta Information page in the PHP docs or the rate limits guide on the Anthropic docs.