[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC):2025-09-02。"],[],[],null,["# Quotas and limits\n\nThis document contains the current API restrictions and usage limits for\nSpeech-to-Text. This page will be updated to reflect any changes to these\nrestrictions and usage limits. We reserve the right to change these limits.\n\nYou can\n[request a quota increase](/docs/quotas/help/request_increase)\nif necessary. See the Google Cloud [quota page](/docs/quota) for more information on\nviewing and managing your quota.\n\nAfter submitting your request, Google might contact you for more\ninformation, and inform you whether your request is approved or denied.\n| **Note:** The time limits listed below refer to the length of the audio file, regardless of whether it is [multichannel](/speech-to-text/v2/docs/multi-channel) audio. Be aware, however, that each channel is billed individually. This can lead to situations where you are billed for more time than than the duration of the audio file. See the [pricing page](/speech-to-text/pricing) for more information.\n\nContent limits\n--------------\n\n### Synchronous requests\n\nSynchronous recognition requests (using the [`Recognize`](/speech-to-text/v2/docs/reference/rpc/google.cloud.speech.v2#google.cloud.speech.v2.Speech.Recognize)\nmethod) accept audio data either inline in the\n[`content`](/speech-to-text/v2/docs/reference/rpc/google.cloud.speech.v2#google.cloud.speech.v2.RecognizeRequest.FIELDS.bytes.google.cloud.speech.v2.RecognizeRequest.content) field of the request or as a\n[Cloud Storage URI](/storage/docs/reference-uris) in the [`uri`](/speech-to-text/v2/docs/reference/rpc/google.cloud.speech.v2#google.cloud.speech.v2.RecognizeRequest.FIELDS.string.google.cloud.speech.v2.RecognizeRequest.uri) field\nof the request. Audio sent to a synchronous request is limited to 10 MB or 1\nminute of audio duration (whichever is reached first). For more information on\nsynchronous recognition, see the\n[synchronous recognition overview](/speech-to-text/v2/docs/sync-recognize).\n\n### Streaming requests\n\nStreaming recognition requests (using the\n[`StreamingRecognize`](/speech-to-text/v2/docs/reference/rpc/google.cloud.speech.v2#google.cloud.speech.v2.Speech.StreamingRecognize) method) only\naccept inline audio in the [`audio`](/speech-to-text/v2/docs/reference/rpc/google.cloud.speech.v2#google.cloud.speech.v2.StreamingRecognizeRequest.FIELDS.bytes.google.cloud.speech.v2.StreamingRecognizeRequest.audio) field of the\nrequest. Each request in the stream is limited to 25 KB of audio. A stream can\nremain open for up to 5 minutes, and the audio must be sent at a rate that\napproximates real time. If you need to stream content for longer than 5 minutes,\nsee the\n[endless streaming tutorial](/speech-to-text/docs/endless-streaming-tutorial). For more information on\nstreaming recognition, see the\n[streaming recognition overview](/speech-to-text/v2/docs/streaming-recognize).\n\n### Batch requests\n\nBatch recognition requests (using the [`BatchRecognize`](/speech-to-text/v2/docs/reference/rpc/google.cloud.speech.v2#google.cloud.speech.v2.Speech.BatchRecognize)\nmethod) only accept audio as a [Cloud Storage URI](/storage/docs/reference-uris)\nin the [`uri`](/speech-to-text/v2/docs/reference/rpc/google.cloud.speech.v2#google.cloud.speech.v2.BatchRecognizeFileMetadata.FIELDS.string.google.cloud.speech.v2.BatchRecognizeFileMetadata.uri) field of the request. Each\n[`BatchRecognizeRequest`](/speech-to-text/v2/docs/reference/rpc/google.cloud.speech.v2#batchrecognizerequest) can contain up to 15\n[`files`](/speech-to-text/v2/docs/reference/rpc/google.cloud.speech.v2#google.cloud.speech.v2.BatchRecognizeRequest.FIELDS.repeated.google.cloud.speech.v2.BatchRecognizeFileMetadata.google.cloud.speech.v2.BatchRecognizeRequest.files) to transcribe. Each file can be up to 8 hours\nin duration. For more information on asynchronous recognition, see the\n[batch recognition overview](/speech-to-text/v2/docs/batch-recognize).\n\n### Multiple language recognition\n\nMultiple language recognition is only available in the global, US, and EU\nSpeech-to-Text endpoints.\n\n### Adaptation\n\nWithin any request, you may also supply\n[PhraseSet and CustomClass](/speech-to-text/v2/docs/adaptation-model) resources.\nThe following limits apply to these resources:\n\nResource limits\n---------------\n\nThe current API resource limits for Speech-to-Text are as follows\n(and are subject to change):\n\nRequest limits\n--------------\n\nThe current API usage limits for Speech-to-Text are as follows\n(and are subject to change):\n\n\\* Streaming recognition has a quota limit of 300 concurrent sessions per\n5 minutes and a limit of 3,000 requests per minute, which applies to\nall concurrent sessions together. The initial configuration request for a\nsession does not count against the request quota.\n\nThese limits apply to each Speech-to-Text developer project, and are shared\nacross all applications and IP addresses using a given a developer project."]]