> ## Documentation Index
> Fetch the complete documentation index at: https://docs.trueparser.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Result Persistence

> Managing document storage and retention.

## Result Persistence

TrueParser provides reliable persistence for your parsed results, ensuring that structured data is available for retrieval by your applications.

### Tenant-Scoped Storage

TrueParser stores all raw uploads and final parsed artifacts in secure, tenant-isolated buckets. By default, the platform provides managed storage, but for production workloads, you can configure your own **S3-compatible storage backend** in the dashboard.

#### The Proxy Access Model

Even when using your own storage, **TrueParser remains the sole client for result access.** This design ensures a consistent, secure, and stable integration contract:

1. **Server-Side Management**: TrueParser uses your provided S3 credentials server-side to write and read artifacts on your behalf.
2. **No Direct Access**: Clients never receive direct S3 URLs, bucket names, or presigned objects. This allows the platform to manage authorization, retention, and scaling without impacting your code.
3. **API Stability**: All result retrieval happens through the TrueParser API, decoupling your application from the underlying storage infrastructure.

***

## Retention Policy

TrueParser is an **operational orchestration platform**, designed to be a high-performance pass-through for your document data.

### 1. Managed Storage (Default)

In the default managed environment, results are retained for **3 hours**. After this window, the jobs and their results are automatically purged from the system.

### 2. Tenant-Specific S3

If you have configured a tenant-scoped S3 bucket, TrueParser uploads the final artifacts directly to your infrastructure. In this mode, you have full control over data retention policies (e.g., Lifecycle Rules) on your own bucket.

### Why 3 Hours?

Short-lived retention allows TrueParser to remain a stateless, high-performance "pass-through" for your data. It is intended that your application:

1. Submits a document.
2. Polls for completion.
3. Downloads the structured JSON.
4. Persists the result into your own specialized storage or database.

***

## Document Identifiers & Overwrites

If you supply a `documentId` during ingestion, TrueParser treats it as the target key for that retention window. If you submit a new document with the same `documentId` within 3 hours, the existing artifact will be overwritten. This allows for simple idempotency or re-parsing workflows without manual deletion.

After the 3-hour window expires, the `documentId` becomes available for reuse.