Name: ByteDance Dolphin API: PDF and Document Parsing to Markdown
Brand: magicapi
SKU: api-magicapi-document-image-parsing-api-bytedance-dolphin
Availability: InStock
Author: API Market

Shashank Agarwal

# 📄 Document Parsing API (ByteDance Dolphin)

## Overview
Dolphin is a document image parsing model from ByteDance. Its developer describes it as "Document Image Parsing via Heterogeneous Anchor Prompting".

Send a PDF or image in `file` and choose `output_format`: `markdown_content` (default) returns the document as Markdown text, `json_content` returns JSON. Headings, paragraphs and lists come back as Markdown structure, as in the example below.

## Features
- **PDF and image input**: send a PDF or an image of a document in `file`.
- **Markdown output**: `markdown_content` (default) returns headings, paragraphs and lists as Markdown.
- **JSON output**: `json_content` returns the parsed content as JSON.
- **Two fields only**: a file URL and an output format.

## Use Cases
1. **PDF to Markdown**: turn reports, papers and articles into Markdown for wikis, search or editing.
2. **Preparing documents for LLMs**: convert PDFs to text before summarizing, chunking or indexing them.
3. **Digitizing scanned pages**: extract the text and structure of photographed or scanned documents.
4. **Document pipelines**: feed parsed content into downstream extraction and classification steps.

## Visual Examples
### Example Input
The first page of the LLaMA research paper, sent as an image and parsed to Markdown.

![First page of the LLaMA paper sent to the Dolphin example request](https://blog.api.market/wp-content/uploads/2026/10/dolphin-input-api-market.jpg)

```json
{
  "file": "https://blog.api.market/wp-content/uploads/2026/10/dolphin-input-api-market.jpg",
  "output_format": "markdown_content"
}
```

### Example Output
The model returned Markdown. The start of the output (shortened):

```markdown
# LLaMA: Open and Efficient Foundation Language Models

Hugo Touvron; Thibaut Lavril; Gautier Izacard; Xavier Martinet Marie-Anne Lachaux, Timothee Lacroix, Baptiste Rozière, Naman Goyal Eric Hambro, Faisal Azhar, Aurelien Rodriguez, Armand Joulin Edouard Grave; Guillaume Lample*

Meta A

## Abstract

We introduce LLaMA, a collection of foundation language models ranging from 7B to 65B parameters. We train our models on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets. In particular, LLaMA-13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA65B is competitive with the best models, Chinchilla-70B and PaLM-540B. We release all our models to the research community $^1$ .


```

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `file` | string (URL) | Yes |  |  | URL of the PDF or image file to parse |
| `output_format` | string | No | `markdown_content` | `markdown_content`, `json_content` | Output format: markdown_content returns Markdown text, json_content returns JSON |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.007 per API unit.
- **1 API unit = 1 request with up to 28 seconds of processing time.** A request that takes longer is billed one unit per started 28-second block (for example, 60 seconds of processing is 3 units, $0.021).
- Processing time grows with the number and density of pages, so larger documents can use several units.
- **Example cost**: the example page took 10.5 seconds of processing, within one 28-second block: 1 API unit = **$0.007**.
- **Failed or cancelled predictions are not billed.**
- **GET requests to check prediction status** are free and don't count toward API unit usage.
- **No monthly subscription fee.**

## API Documentation

### Base URL
All API requests go to: `https://prod.api.market/api/v1/magicapi/document-image-parsing-api-bytedance-dolphin`

### Authentication
Send your api.market API key in the `x-api-market-key` header:
```
x-api-market-key: YOUR_API_MARKET_KEY
```

### Asynchronous Processing
**IMPORTANT**: This API is asynchronous. The flow is:
1. Send a POST request to `/predictions`. A successful request returns **201 Created** with a prediction `id` and a `status` of `starting` or `processing`.
2. Poll GET `/predictions/{id}` with that ID.
3. Keep polling while `status` is `starting` or `processing`.
4. Stop when `status` is `succeeded` (read `output`), `failed` or `canceled` (read `error`).

A polling interval of 2 seconds and an overall timeout of 900 seconds (15 minutes) work well. Parsing a long PDF can take several minutes; the one-page example took about 10.5 seconds.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/document-image-parsing-api-bytedance-dolphin/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "3076d1f489b2365a2cc7d20d440a9a93e469a1eb145d85a7913a5c920b26d126",
    "input": {
      "file": "https://blog.api.market/wp-content/uploads/2026/10/dolphin-input-api-market.jpg",
      "output_format": "markdown_content"
    }
  }'
```

Response (201 Created):
```json
{
  "id": "tehtdyyff1rgc0d1cdhrp98wm4",
  "version": "3076d1f489b2365a2cc7d20d440a9a93e469a1eb145d85a7913a5c920b26d126",
  "status": "starting",
  "created_at": "2026-10-22T11:34:08.140Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/document-image-parsing-api-bytedance-dolphin/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "tehtdyyff1rgc0d1cdhrp98wm4",
  "status": "succeeded",
  "output": "# LLaMA: Open and Efficient Foundation Language Models\n\nHugo Touvron; Thibaut Lavril; Gautier Izacard; Xavier Martinet Marie-Anne Lachaux, Timothee Lacroix, Baptiste Rozière, Naman Goyal Eric Hambro, Faisal Azhar, Aurelien Rodriguez, Armand Joulin Edouard Grave; Guillaume Lample*\n\nMeta A\n\n## Abstract\n\nWe introduce LLaMA, a collection of foundation language models ranging from 7B to 65B parameters. We train our models on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets. In particular, LLaMA-13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA65B is competitive with the best models, Chinchilla-70B and PaLM-540B. We release all our models to the research community $^1$ .\n\n...",
  "error": null,
  "created_at": "2026-10-22T11:34:08.140Z",
  "started_at": "2026-10-22T11:36:03.915Z",
  "completed_at": "2026-10-22T11:36:14.447Z"
}
```

With `markdown_content`, `output` is a single string of Markdown (shortened above). With `json_content`, it holds the parsed content as JSON.

### Status Codes
- **201 Created**: prediction created (POST `/predictions`). This is the normal success response.
- **200 OK**: prediction details returned (GET `/predictions/{id}`).
- **400 Bad Request**: invalid input, for example a missing `file` or an `output_format` other than `markdown_content` or `json_content`.
- **401 Unauthorized**: missing or invalid API key.
- **404 Not Found**: no prediction with that ID.
- **429 Too Many Requests**: rate limit exceeded.
- **500 Internal Server Error** or **503 Service Unavailable**: temporary server problem; retry later.

## Rate Limits
All plans allow 10 requests per second. Requests above that return `429 Too Many Requests`; back off before retrying and honour any `Retry-After` header. Polling a prediction with GET is free but still counts toward the per-second limit, so poll every 1 to 5 seconds rather than in a tight loop.

## Error Handling
- Check the HTTP status code of every request and read the response body for the error message.
- A prediction can also end with `status` set to `failed`; the `error` field then explains why. Failed predictions are not billed.
- Retry 429, 500 and 503 responses with exponential backoff. Fix the input before retrying a 400.

## Implementation Guide
### Example Code

#### JavaScript
```javascript
const apiKey = "YOUR_API_MARKET_KEY";
const baseUrl = "https://prod.api.market/api/v1/magicapi/document-image-parsing-api-bytedance-dolphin";

async function createAndWaitForPrediction(input) {
  const response = await fetch(`${baseUrl}/predictions`, {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      "x-api-market-key": apiKey
    },
    body: JSON.stringify({ version: "3076d1f489b2365a2cc7d20d440a9a93e469a1eb145d85a7913a5c920b26d126", input })
  });

  if (response.status !== 201) {
    throw new Error(`API error: ${response.status} ${await response.text()}`);
  }

  const prediction = await response.json();
  return pollForResults(prediction.id);
}

const delay = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

async function pollForResults(predictionId) {
  const maxTimeout = 900000;
  const pollInterval = 2000;
  const startTime = Date.now();

  while (Date.now() - startTime < maxTimeout) {
    const response = await fetch(`${baseUrl}/predictions/${predictionId}`, {
      headers: { "x-api-market-key": apiKey }
    });

    if (!response.ok) {
      throw new Error(`API error: ${response.status} ${await response.text()}`);
    }

    const prediction = await response.json();

    if (prediction.status === "succeeded") {
      return prediction.output;
    }

    if (prediction.status === "failed" || prediction.status === "canceled") {
      throw new Error(`Prediction ${prediction.status}: ${prediction.error || "Unknown error"}`);
    }

    await delay(pollInterval);
  }

  throw new Error("Prediction timed out after 15 minutes");
}

createAndWaitForPrediction({
    "file": "https://blog.api.market/wp-content/uploads/2026/10/dolphin-input-api-market.jpg",
    "output_format": "markdown_content"
  }).then((output) => console.log(output));
```

#### Python
```python
import time
import requests

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/document-image-parsing-api-bytedance-dolphin"


def create_and_wait_for_prediction(input_data):
    headers = {"Content-Type": "application/json", "x-api-market-key": API_KEY}
    response = requests.post(
        f"{BASE_URL}/predictions",
        headers=headers,
        json={"version": "3076d1f489b2365a2cc7d20d440a9a93e469a1eb145d85a7913a5c920b26d126", "input": input_data},
    )

    if response.status_code != 201:
        raise Exception(f"API error: {response.status_code} {response.text}")

    return poll_for_results(response.json()["id"])


def poll_for_results(prediction_id):
    max_timeout = 900
    poll_interval = 2
    headers = {"x-api-market-key": API_KEY}
    start_time = time.time()

    while time.time() - start_time < max_timeout:
        response = requests.get(f"{BASE_URL}/predictions/{prediction_id}", headers=headers)

        if response.status_code != 200:
            raise Exception(f"API error: {response.status_code} {response.text}")

        prediction = response.json()

        if prediction["status"] == "succeeded":
            return prediction["output"]

        if prediction["status"] in ("failed", "canceled"):
            raise Exception(f"Prediction {prediction['status']}: {prediction.get('error') or 'Unknown error'}")

        time.sleep(poll_interval)

    raise Exception("Prediction timed out after 15 minutes")


print(create_and_wait_for_prediction({
    "file": "https://blog.api.market/wp-content/uploads/2026/10/dolphin-input-api-market.jpg",
    "output_format": "markdown_content"
}))
```

## FAQ
- **Which files can I send?**
  - A PDF or an image of a document, as a URL in `file`.
- **What is the difference between the two output formats?**
  - `markdown_content` (default) returns the document as Markdown text. `json_content` returns the parsed content as JSON.
- **How long does a request take?**
  - It depends on the document. The example page took about 10.5 seconds of processing; multi-page PDFs take longer, so poll for several minutes before giving up.
- **How much does a document cost?**
  - $0.007 per started 28 seconds of processing. The example page took 10.5 seconds, which is 1 unit ($0.007).
- **What if my prediction fails?**
  - Failed predictions are not billed. Read the `error` field, adjust the input and try again.

## Contact Us
If you have any questions, reach out to us through the chat on the bottom right of the page. You can also email our support team at hello@magicapi.com.


Parse PDFs and document images into Markdown or JSON with ByteDance's Dolphin document image parsing model. $0.007 per request unit, pay as you go.