---
title: "Create download link"
description: "A short-lived URL that downloads the layer's current PDF without an Authorization header, for services that only take a plain URL."
source: "https://www.cloudpdf.com/docs/api-reference/document-operations/download-link"
---

# Create download link

`POST /v1/docs/{docId}/layers/{layerName}/download-link` · v3.3.0

A short-lived URL that downloads the layer's current PDF without an Authorization header, for services that only take a plain URL.

The link opens the layer as it is when created: once the layer changes, it answers 404, so create it right before handing it over. It downloads that one file and nothing else, and a password-protected document is checked when the link is created. Deployments that verify tokens with their own keys (JWKS) mint links on their backend: a doc token with \`layer\_name\`, a \`download: \{ docVersion, mode: "incremental" }\` claim and a short \`exp\`, passed as \`link\` on the versioned download URL.

## Authentication

Any one of these credentials is accepted.

- **apiToken** — The deployment's static root credential (CLOUDPDF\_API\_AUTH\_TOKENS), valid on every surface.
- **tenantToken** — Delegated tenant JWT, valid only for its own tenant: under /v1/tenants/\{tenantId}/ the path tenant must equal the token's tenant\_id, and on /v1/docs/\{docId} the document must belong to that tenant and the token's scope must hold the operation's document permissions, as a document token does. For servers; browsers get document tokens. Doc-scoped viewer tokens are rejected on every admin route.
- **docToken** — Doc-scoped JWT, valid only on the /v1/docs/\{docId} subtree it names, gated by the capability scopes it carries (each operation's x-required-capability).

**Document capability:** `doc.download`

## Parameters

- `docId` string (path) — required
- `layerName` string (path) — required
- `X-Document-Password` string (header) — Base64-encoded password for an encrypted document. Valid only with the API token (403 anywhere else). An encrypted document answers 422 DocPasswordRequired when the header is absent. Viewer doc JWTs use the SDK password-session flow instead.

## Request body

`application/json` — DocDownloadLinkRequest

- `expiresInSeconds` integer

**Example**

```json
{
  "expiresInSeconds": 1
}
```

## SDK examples

**TypeScript**

```typescript
import { CloudPDFClient } from "@cloudpdf/sdk";

const client = new CloudPDFClient({
  token: "<token>",
});

await client.doc.downloadLink({
  docId: "docId",
  layerName: "layerName"
});
```

**Python**

```python
from cloudpdf import CloudPDFClient

client = CloudPDFClient(
    token="<token>",
)

client.doc.download_link(
    doc_id="docId",
    layer_name="layerName",
)
```

**PHP**

```php
use CloudPDF\CloudPDFClient;

$client = new CloudPDFClient(
    token: '<token>',
);

$client->doc->downloadLink(
    'docId',
    'layerName',
    new DocDownloadLinkRequest([]),
);
```

**.NET**

```csharp
using CloudPDF;

var client = new CloudPDFClient(
    "<token>"
);

await client.Doc.DownloadLinkAsync(
    new DocDownloadLinkRequest { DocId = "docId", LayerName = "layerName" }
);
```

**Go**

```go
import (
    "context"

    cloudpdf "github.com/embedpdf/cloudpdf-sdk-go/v3"
    client "github.com/embedpdf/cloudpdf-sdk-go/v3/client"
    option "github.com/embedpdf/cloudpdf-sdk-go/v3/option"
)

client := client.NewClient(
    option.WithToken("<token>"),
)

request := &cloudpdf.DocDownloadLinkRequest{
    DocID: "docId",
    LayerName: "layerName",
}
client.Doc.DownloadLink(
    context.TODO(),
    request,
)
```

**Java**

```java
import com.cloudpdf.api.CloudPDFClient;

CloudPDFClient client = CloudPDFClient
    .builder()
    .token("<token>")
    .build();

client.doc().downloadLink(
    "docId",
    "layerName",
    DocDownloadLinkRequest
        .builder()
        .build()
);
```

**Ruby**

```ruby
require "cloudpdf"

client = CloudPDF::Client.new(
  token: "<token>"
)

client.doc.download_link(
  doc_id: "docId",
  layer_name: "layerName"
)
```

## Responses

### 200 — OK

`application/json` — DocDownloadLinkResponse

- `url` string — required
- `expiresAt` string\<date-time> — required

### 404 — Not found

`application/json` — ErrorResponse

- `error` EngineError — required
  - `code` "Unknown" | "InvalidArg" | "DocNotOpen" | "DocOpenFailed" | "DocPasswordRequired" | "DocPasswordIncorrect" | "GrantPassphraseRequired" | "Aborted" | "Network" | "Unauthenticated" | "Forbidden" | "NotFound" | "WireFormat" | "RuntimeUnavailable" | "LayerVersionConflict" | "NotImplemented" | "MalformedPdf" | "SigningPending" | "SigningExpired" | "SigningVersionMismatch" | "SignatureRefused" | "ProtectedDocument" | "PayloadTooLarge" | "ObjectNumberUnavailable" | "LayerFull" | "ChangeConflict" | "UndoUnavailable" | "IdempotencyKeyReused" | "LayerDeleted" | "VisitorIdentityRequired" | "Conflict" | "EngineBusy" | "EngineRestarting" | "DocumentQuarantined" | "ThumbnailNotReady" | "TooManyRequests" | "GrantExpired" | "OriginRequired" | "OriginNotAllowed" | "FeatureNotEnabled" | "FeatureStopped" | "LicenseInvalid" | "LicenseUsageLimit" | "UpstreamError" — required — Why the request failed. A newer server can add codes: read one your client doesn't know as \`Unknown\`.
  - `message` string — required
  - `details` object — What a code carries besides its message, as each code says.

### 501 — Not implemented

`application/json` — ErrorResponse

- `error` EngineError — required
  - `code` "Unknown" | "InvalidArg" | "DocNotOpen" | "DocOpenFailed" | "DocPasswordRequired" | "DocPasswordIncorrect" | "GrantPassphraseRequired" | "Aborted" | "Network" | "Unauthenticated" | "Forbidden" | "NotFound" | "WireFormat" | "RuntimeUnavailable" | "LayerVersionConflict" | "NotImplemented" | "MalformedPdf" | "SigningPending" | "SigningExpired" | "SigningVersionMismatch" | "SignatureRefused" | "ProtectedDocument" | "PayloadTooLarge" | "ObjectNumberUnavailable" | "LayerFull" | "ChangeConflict" | "UndoUnavailable" | "IdempotencyKeyReused" | "LayerDeleted" | "VisitorIdentityRequired" | "Conflict" | "EngineBusy" | "EngineRestarting" | "DocumentQuarantined" | "ThumbnailNotReady" | "TooManyRequests" | "GrantExpired" | "OriginRequired" | "OriginNotAllowed" | "FeatureNotEnabled" | "FeatureStopped" | "LicenseInvalid" | "LicenseUsageLimit" | "UpstreamError" — required — Why the request failed. A newer server can add codes: read one your client doesn't know as \`Unknown\`.
  - `message` string — required
  - `details` object — What a code carries besides its message, as each code says.

### default — Error

`application/json` — ErrorResponse

- `error` EngineError — required
  - `code` "Unknown" | "InvalidArg" | "DocNotOpen" | "DocOpenFailed" | "DocPasswordRequired" | "DocPasswordIncorrect" | "GrantPassphraseRequired" | "Aborted" | "Network" | "Unauthenticated" | "Forbidden" | "NotFound" | "WireFormat" | "RuntimeUnavailable" | "LayerVersionConflict" | "NotImplemented" | "MalformedPdf" | "SigningPending" | "SigningExpired" | "SigningVersionMismatch" | "SignatureRefused" | "ProtectedDocument" | "PayloadTooLarge" | "ObjectNumberUnavailable" | "LayerFull" | "ChangeConflict" | "UndoUnavailable" | "IdempotencyKeyReused" | "LayerDeleted" | "VisitorIdentityRequired" | "Conflict" | "EngineBusy" | "EngineRestarting" | "DocumentQuarantined" | "ThumbnailNotReady" | "TooManyRequests" | "GrantExpired" | "OriginRequired" | "OriginNotAllowed" | "FeatureNotEnabled" | "FeatureStopped" | "LicenseInvalid" | "LicenseUsageLimit" | "UpstreamError" — required — Why the request failed. A newer server can add codes: read one your client doesn't know as \`Unknown\`.
  - `message` string — required
  - `details` object — What a code carries besides its message, as each code says.
