---
title: PDF proxy
description: Fetch the original scanned PDF backing an indexed document.
order: 5
---

`GET /pdf/{container}/{file}`

Proxies the original PDF from blob storage — no API key required, this route is unauthenticated and IP-rate-limited instead (30 req/min).

`container` and `file` are validated server-side: `container` must be alphanumeric (plus `-`/`_`); `file` must be alphanumeric (plus `-`/`_`) and end in `.pdf`. Values outside that shape are rejected before any lookup happens, not passed through to storage.

```bash
curl https://api.dike.it.com/pdf/legislation/eg-leg-151912.pdf -o eg-leg-151912.pdf
```

The `container`/`file` pair for a given document isn't returned by `/v1/search` or `/v1/document` today — construct it from the document's `canonical_id` and known container name, or treat this endpoint as best-effort until a `source` field is exposed on those responses.

## Errors

- `404 not_found` — file doesn't exist in storage
- `422 validation_error` — `container`/`file` failed the shape check above