Try Gemini 1.5 models, the latest multimodal models in Vertex AI, and see what you can build with up to a 2M token context window. Try Gemini 1.5 models, the latest multimodal models in Vertex AI, and see what you can build with up to a 2M token context window.

Quotas and limits

This document lists the quotas and limits that apply to Speech-to-Text.

Google Cloud uses quotas to help ensure fairness and reduce spikes in resource use and availability. A quota restricts how much of a Google Cloud resource your Google Cloud project can use. Quotas apply to a range of resource types, including hardware, software, and network components. For example, quotas can restrict the number of API calls to a service, the number of load balancers used concurrently by your project, or the number of projects that you can create. Quotas protect the community of Google Cloud users by preventing the overloading of services. Quotas also help you to manage your own Google Cloud resources.

The Cloud Quotas system does the following:

Monitors your consumption of Google Cloud products and services
Restricts your consumption of those resources
Provides a means to request changes to the quota value

In most cases, when you attempt to consume more of a resource than its quota allows, the system blocks access to the resource, and the task that you're trying to perform fails.

Quotas generally apply at the Google Cloud project level. Your use of a resource in one project doesn't affect your available quota in another project. Within a Google Cloud project, quotas are shared across all applications and IP addresses.

To increase or decrease most quotas, use the Google Cloud console. For more information, see Request a higher quota.

There are also limits on Speech-to-Text resources. These limits are unrelated to the quota system. Limits cannot be changed unless otherwise stated.

This document contains the current API restrictions and usage limits for Cloud Speech-to-Text. This page will be updated to reflect any changes to these restrictions and usage limits. We reserve the right to change these limits.

Content Limits

Content to Speech-to-Text is provided as audio data, either directly within the content field of the request or referenced within a Google Cloud Storage URI in the uri field of the request. There is a limit of 10 MB on all single requests sent to the API using local files. In the case of the Recognize and LongRunningRecognize methods, this limit applies to the size of the request sent. In the case of the StreamingRecognize method, the 10 MB limit applies to both the initial StreamingRecognize request and the size of each individual message in the stream. Exceeding this limit will throw an error. There is no size limit on requests sent using audio data stored in a Google Cloud Storage bucket.

The API contains the following limits on the size of this content (and are subject to change).

Content Limit	Audio Length
Synchronous Requests	~1 Minute
Asynchronous Requests	~480 Minutes*
Streaming Requests	~5 Minutes**

* Audio longer than ~1 minute must use the uri field to reference an audio file in Google Cloud Storage.
** If you need to stream content for more than 5 minutes, see the endless streaming tutorial.

For StreamingRecognize requests, audio must be sent at a rate that approximates real time.

Attempting to process content in excess of these content limits will produce an error. For more information, see Error messages and Troubleshooting.

Within any request, you may also supply a PhraseSet resource containing a list of phrases specific to the request. (A single word counts as a phrase in this context.) The following limits apply to such a context:

Speech Adaptation Limit	Value
Phrases per request	5000
Total characters per request	100,000
Characters per phrase	100

Request Limits

The current API usage limits for Speech-to-Text are as follows (and are subject to change):

Type of Limit	Usage Limit
Recognition Requests per 60 seconds*	900
Adaptation resource requests per 60 seconds*	10
Processing per day	480 hours of audio

* Each StreamingRecognize session is considered a single request even though it includes multiple frames of StreamingRecognizeRequest audio within the stream.

Requests and/or attempts at audio processing in excess of these limits will produce an error. For more information, see Error messages and Troubleshooting.

These limits apply to each Speech-to-Text developer project, and are shared across all applications and IP addresses using a given a developer project.