REST API Guide

Guide

Parse

Guides

Parse REST API endpoints, request/response shapes, expand values, and presigned URL patterns.

For the complete SDK and API documentation, see the Parse API Reference →

This page covers REST API usage patterns: endpoint overview, request examples, response shapes, and the expand parameter quick-reference.

Endpoints

Method	URL	Use case
`POST /api/v2/parse`	Parse a file by ID or URL (JSON body)
`POST /api/v2/parse/upload`	Upload a file and parse it in one request (multipart/form-data)
`GET /api/v2/parse`	List and filter parse jobs with pagination
`GET /api/v2/parse/{job_id}`	Check job status and retrieve results

Required header: Authorization: Bearer YOUR_API_KEY on all endpoints.

Optional header (Enterprise): Usage-Tags attributes a request’s usage to a team, environment, or customer for billing breakdowns. See Usage Tags.

Creating a parse job

Send a JSON body with file_id (or source_url), tier, version, and any options:

curl -X POST 'https://api.cloud.llamaindex.ai/api/v2/parse' \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY" \
  --data '{
    "file_id": "<file_id>",
    "tier": "agentic",
    "version": "latest"
  }'

The create response returns the job fields at the top level — id, project_id, status, and tier (there is no job wrapper here). Use the top-level id to retrieve results with the expand query parameter (see Retrieving results below). See Configuring Parse for every available option.

Options like page_ranges, crop_box, and output_options nest inside the same JSON body — they are not top-level fields. A bare "max_pages": 2 alongside file_id/tier/version returns 422 extra_forbidden; nest it under page_ranges instead:

curl -X POST 'https://api.cloud.llamaindex.ai/api/v2/parse' \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY" \
  --data '{
    "file_id": "<file_id>",
    "tier": "agentic",
    "version": "latest",
    "page_ranges": {"max_pages": 2}
  }'

See Configuring Parse for every available option.

Uploading and parsing a file

If you don’t already have a file_id, POST /api/v2/parse/upload uploads a file and starts a parse job in a single multipart/form-data request. Send the file as a file field and the parse configuration as a configuration JSON string field:

curl -X POST 'https://api.cloud.llamaindex.ai/api/v2/parse/upload' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY" \
  -F 'file=@/path/to/your/file.pdf' \
  -F 'configuration={"tier": "agentic", "version": "latest"}'

Like the JSON create endpoint above, the response returns id and status at the top level. See Configuring Parse for every field you can set inside configuration.

Listing jobs

curl 'https://api.cloud.llamaindex.ai/api/v2/parse?page_size=10&status=COMPLETED' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY"

Query parameters:

Parameter	Type	Description
`page_size`	integer	Max items per page (optional)
`page_token`	string	Token for the next page (from a previous response)
`status`	string	Filter: `PENDING`, `RUNNING`, `COMPLETED`, `FAILED`, `CANCELLED`

Response:

{
  "items": [
    {
      "id": "job-uuid-1",
      "project_id": "project-uuid",
      "status": "COMPLETED",
      "error_message": null
    }
  ],
  "next_page_token": "eyJsYXN0X2lkIjogImpvYi...",
  "total_size": 42
}

Retrieving results

curl 'https://api.cloud.llamaindex.ai/api/v2/parse/{job_id}?expand=markdown,items,metadata' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY"

The response always includes a job object with id, project_id, status, and error_message. Additional fields depend on the expand parameter.

Expand values

Value	What it returns
`text`	Plain text per page
`markdown`	Markdown per page
`items`	Structured JSON items (tables, headings, paragraphs per page)
`metadata`	Page-level metadata (confidence, speaker notes, cost_optimized, etc.)
`job_metadata`	Usage and processing details
`text_full`	Full plain text as a single string (all pages concatenated)
`markdown_full`	Full markdown as a single string (all pages concatenated)
`text_content_metadata`	Text file presigned download URL
`markdown_content_metadata`	Markdown file presigned download URL
`items_content_metadata`	Items file presigned download URL
`metadata_content_metadata`	Metadata file presigned download URL
`text_full_content_metadata`	Full text file presigned download URL
`markdown_full_content_metadata`	Full markdown file presigned download URL
`xlsx_content_metadata`	XLSX file presigned download URL
`output_pdf_content_metadata`	Output PDF presigned download URL
`images_content_metadata`	Images metadata with per-image presigned URLs

Combine multiple values: ?expand=markdown,items,images_content_metadata

For detailed guidance on choosing expand values, see Retrieving Results.

Content metadata response shapes

When requesting *_content_metadata expand values, the response includes presigned URLs for direct download:

XLSX and PDF:

{
  "result_content_metadata": {
    "xlsx": {
      "size_bytes": 15234,
      "exists": true,
      "presigned_url": "https://s3.amazonaws.com/..."
    },
    "outputPDF": {
      "size_bytes": 102400,
      "exists": true,
      "presigned_url": "https://s3.amazonaws.com/..."
    }
  }
}

Images:

{
  "images_content_metadata": {
    "total_count": 3,
    "images": [
      {
        "index": 0,
        "filename": "image_0.png",
        "content_type": "image/png",
        "size_bytes": 12345,
        "presigned_url": "https://s3.amazonaws.com/..."
      }
    ]
  }
}

You can filter images with the image_filenames query parameter:

curl 'https://api.cloud.llamaindex.ai/api/v2/parse/{job_id}?expand=images_content_metadata&image_filenames=image_0.png,image_1.jpg' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY"

Presigned URLs are temporary. Download files promptly after retrieving them, or call the endpoint again for fresh URLs.

Error responses

v2 returns structured validation errors:

{
  "detail": [
    {
      "type": "value_error",
      "loc": ["tier"],
      "msg": "Unsupported tier: invalid_tier. Must be one of: fast, cost_effective, agentic, agentic_plus",
      "input": {}
    }
  ]
}