FeaturesMultimodal

Documents

Send PDFs and documents to document-understanding models.

Gateway supports sending PDF documents to models with document input capability. Pass documents as base64-encoded data or a URL — Gateway handles the provider-specific format translation.

Send a document (base64)

1import base64
2
3with open("report.pdf", "rb") as f:
4    pdf_data = base64.standard_b64encode(f.read()).decode("utf-8")
5
6response = client.responses.create(
7    model="anthropic/claude-sonnet-4-20250514",
8    input=[
9        {
10            "type": "message",
11            "role": "user",
12            "content": [
13                {"type": "text", "text": "Summarize this document."},
14                {
15                    "type": "document",
16                    "source_type": "base64",
17                    "media_type": "application/pdf",
18                    "data": pdf_data,
19                },
20            ],
21        }
22    ],
23)
24
25print(response.output[0].content[0].text)

Send a document (URL)

URL
1response = client.responses.create(
2    model="google/gemini-2.0-flash",
3    input=[
4        {
5            "type": "message",
6            "role": "user",
7            "content": [
8                {"type": "text", "text": "What are the key findings?"},
9                {
10                    "type": "document",
11                    "source_type": "url",
12                    "media_type": "application/pdf",
13                    "data": "https://example.com/report.pdf",
14                },
15            ],
16        }
17    ],
18)

Supported models

Document input is supported by models with "document" in their capabilities.input array. Check model capabilities via GET /v1/models.

ProviderModels
AnthropicClaude Sonnet 4, Claude Haiku 3.5
GoogleGemini 2.0 Flash, Gemini 2.5 Pro

Documents are estimated at ~1,500 tokens per page for compression calculations. A 10-page PDF counts as roughly 15,000 tokens toward the context window.