> ## Documentation Index
> Fetch the complete documentation index at: https://gladia-95-fix-geo-data-retention.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.
# Data retention and zero data retention
> Understand Gladia retention windows for audio, transcripts and metadata, Enterprise zero data retention, callback delivery and retrieval limitations.
Gladia's data retention depends on the plan and the type of data. Audio and transcript retention are separate from API metadata and operational logs. Enterprise customers can request custom retention or zero data retention. Under the documented zero-retention workflow, results are delivered through callbacks and are not available for later retrieval through the API or playground. The data types are:
* **Audio input**: Audio files or audio streams provided for transcription
* **Transcription output**: Text, timestamps, words, utterances
* **API Metadata**: Request IDs, timestamps, processing status
* **Logs**: Operational logs for system reliability
The duration for which your data is stored depends on your plan type. You have two main options for data retention:
* **Standard retention:** Audio and transcripts remain retrievable during the default retention window for your plan and data type.
* **Custom retention (Enterprise):** Enterprise customers can request a custom period. A custom value of zero means deletion within 24 hours; it does not mean immediate deletion.
* **Zero data retention (Enterprise):** This separate option uses ephemeral processing and immediate deletion after processing, as described below. It is not the same as a custom retention value of zero.
Only Enterprise users are eligible for custom data retention and the zero data retention option.
## Default retention by plan
### Free
By default, Free accounts retain **all** data types for **1 year**:
| Data type | Default retention |
| - | - |
| Audio input | 1 year |
| Transcription output | 1 year |
| API Metadata | 1 year |
| Logs | 1 year |
### Paid (Starter, Growth, Enterprise)
By default, paid accounts keep audio and transcripts for a shorter window, while essential API metadata is retained longer for usage tracking:
| Data type | Default retention | What is retained |
| - | - | - |
| Audio input | 3 weeks | Audio files or audio streams provided for transcription |
| Transcription output | 3 weeks | Text, timestamps, words, utterances |
| API Metadata | 1 year | Request IDs, timestamps, and processing status only |
| Logs | Limited period | Operational logs for service quality and reliability |
## Zero data retention behavior
When Zero Data Retention is enabled, Gladia processes data ephemerally; no data is stored at rest.
Enabling Zero Data Retention is a **breaking change** if your integration relies on file upload ([`/v2/upload`](/api-reference/v2/upload/audio-file)) or result retrieval via polling ([`GET /v2/pre-recorded/:id`](/api-reference/v2/pre-recorded/get)). Review the restrictions below and update your integration **before** enabling ZDR.
### Disabled endpoints
* **File upload is disabled**: The [`/v2/upload`](/api-reference/v2/upload/audio-file) endpoint is fully disabled. Any request to this endpoint will fail. The asynchronous API must use an external audio file URL (e.g. an S3 presigned URL or any publicly accessible URL) passed directly as `audio_url` to [`POST /v2/pre-recorded`](/api-reference/v2/pre-recorded/init).
* **Result polling is disabled**: Transcription results cannot be retrieved via [`GET /v2/pre-recorded/:id`](/api-reference/v2/pre-recorded/get). The only way to receive results is through [callbacks](/api-reference/v2/pre-recorded/callback/success).
### No data stored
* **No audio files are stored**: Files cannot be retrieved through the API or in [Gladia's playground](https://app.gladia.io/?utm_campaign=197609689-Website\&utm_source=documentation\&utm_content=sign-up).
* **No transcripts are stored**: Transcription results are not visible in the API or in [Gladia's playground](https://app.gladia.io/?utm_campaign=197609689-Website\&utm_source=documentation\&utm_content=sign-up).
* **No metadata retrieval**: Transcription API calls, audio duration, and other metadata cannot be retrieved through the API or in [Gladia's playground](https://app.gladia.io/?utm_campaign=197609689-Website\&utm_source=documentation\&utm_content=sign-up).
Once the result is delivered via callback, the audio, transcript, and metadata cannot be accessed.
### Migration checklist
Before enabling Zero Data Retention, make sure your integration meets the following requirements:
Replace any usage of `/v2/upload` with an external storage provider. Pass a publicly accessible or signed URL (e.g. AWS S3 presigned URL, GCS signed URL) as `audio_url` when creating a transcription job.
Configure a callback URL in your transcription requests to receive results, since polling and the playground will not be available. See [callback configuration](/api-reference/v2/pre-recorded/init) for setup details.
Remove any code that retrieves transcription results via `GET /v2/pre-recorded/:id` or relies on the Gladia playground for debugging. Results are delivered exclusively through callbacks.
For pre-recorded transcription with zero data retention, use an external audio URL and deliver results through a callback. File upload is unavailable under zero data retention, and results cannot be retrieved later through the API or playground. See the [pre-recorded request reference](/api-reference/v2/pre-recorded/init).
For current security attestations, data residency information, and agreements, see the [compliance hub](https://www.gladia.io/compliance-hub). Plan eligibility for custom retention and zero data retention is listed on [Pricing](https://www.gladia.io/pricing). Choosing a retention setting does not by itself establish your application's overall legal compliance.
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.