API 参考

强大、简洁的 RESTful API,几行代码即可实现专业的 PDF 转电子书功能

API 请求

基础 URL:

https://fusion-api.oomol.com/v1

认证

所有请求需要在 Authorization 请求头中携带 API 密钥（格式：Bearer YOUR_API_KEY）

Authorization: Bearer YOUR_API_KEY

POST/pdf-transform-markdown/submit

提交 PDF 转 Markdown 任务

上传 PDF 文件 URL 并提交转换为 Markdown 格式的任务

请求参数

pdfURL - PDF 文件的云端 URL（string，必需，从文件上传接口获取）
model - 转换模型（string，必需，固定值为 "gundam"）
filename - 文件名(不含扩展名,可选),用于用户下载后识别文件
ignore_pdf_errors - 是否忽略 PDF 解析错误(可选,默认为 true)
ignore_ocr_errors - 是否忽略 OCR 错误(可选,默认为 true)

请求示例

{
  "pdfURL": "https://pdfcraft.ai/examples/api-quickstart.pdf",
  "model": "gundam",
  "filename": "my-document",
  "ignore_pdf_errors": true,
  "ignore_ocr_errors": true
}

响应示例

{
  "success": true,
  "sessionID": "019aa097-f28d-7000-8d56-6a2987a7b144"
}

GET/pdf-transform-markdown/result/:taskId

查询 PDF 转 Markdown 结果

根据任务 ID 查询转换任务的状态和结果

URL 参数

taskId - 任务 ID（从提交接口返回的 sessionID）

响应示例（处理中）

{
  "success": true,
  "state": "processing",
  "progress": 66
}

响应示例（已完成）

{
  "success": true,
  "state": "completed",
  "progress": 100,
  "data": {
    "downloadURL": "https://cdn.oomol.com/result.md"
  }
}

响应示例（失败）

{
  "success": false,
  "state": "failed",
  "progress": 0,
  "error": "转换失败原因"
}

POST/pdf-transform-epub/submit

提交 PDF 转 EPUB 任务

上传 PDF 文件 URL 并提交转换为 EPUB 格式的任务

请求参数

pdfURL - PDF 文件的云端 URL（string，必需，从文件上传接口获取）
model - 转换模型（string，必需，固定值为 "gundam"）
filename - 文件名(不含扩展名,可选),用于用户下载后识别文件
ignore_pdf_errors - 是否忽略 PDF 解析错误(可选,默认为 true)
ignore_ocr_errors - 是否忽略 OCR 错误(可选,默认为 true)

请求示例

{
  "pdfURL": "https://pdfcraft.ai/examples/api-quickstart.pdf",
  "model": "gundam",
  "filename": "my-document",
  "ignore_pdf_errors": true,
  "ignore_ocr_errors": true
}

响应示例

{
  "success": true,
  "sessionID": "019aa097-f28d-7000-8d56-6a2987a7b144"
}

GET/pdf-transform-epub/result/:taskId

查询 PDF 转 EPUB 结果

根据任务 ID 查询转换任务的状态和结果

URL 参数

taskId - 任务 ID（从提交接口返回的 sessionID）

响应示例（处理中）

{
  "success": true,
  "state": "processing",
  "progress": 66
}

响应示例（已完成）

{
  "success": true,
  "state": "completed",
  "progress": 100,
  "data": {
    "downloadURL": "https://cdn.oomol.com/result.epub"
  }
}

响应示例（失败）

{
  "success": false,
  "state": "failed",
  "progress": 0,
  "error": "转换失败原因"
}

cURL

# 1. 提交 PDF 转 Markdown 任务
curl -X POST https://fusion-api.oomol.com/v1/pdf-transform-markdown/submit \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "pdfURL": "https://pdfcraft.ai/examples/api-quickstart.pdf",
    "model": "gundam",
    "filename": "my-document"
  }'

# 响应: {"success": true, "sessionID": "019aa097-f28d-7000-8d56-6a2987a7b144"}

# 2. 查询转换结果
curl -X GET https://fusion-api.oomol.com/v1/pdf-transform-markdown/result/019aa097-f28d-7000-8d56-6a2987a7b144 \
  -H "Authorization: Bearer YOUR_API_KEY"

# 响应: {"success": true, "state": "completed", "data": {"downloadURL": "https://cdn.oomol.com/result.md"}}

Python SDK

用于与 PDF Craft API 交互的 Python SDK。它通过处理身份验证、任务提交和结果轮询，简化了将 PDF 转换为 Markdown 或 EPUB 的过程。

在 GitHub 上查看

安装

pip install pdf-craft-sdk

基本用法

from pdf_craft_sdk import PDFCraftClient

# Initialize the client
client = PDFCraftClient(api_key="YOUR_API_KEY")

# Convert a local PDF to Markdown
try:
    download_url = client.convert_local_pdf("document.pdf")
    print(f"Conversion successful! Download URL: {download_url}")
except Exception as e:
    print(f"An error occurred: {e}")

本地文件上传与进度跟踪

from pdf_craft_sdk import PDFCraftClient, UploadProgress

def on_progress(progress: UploadProgress):
    print(f"Upload: {progress.percentage:.2f}% - Part {progress.current_part}/{progress.total_parts}")

client = PDFCraftClient(api_key="YOUR_API_KEY")

# Upload local file with progress tracking
download_url = client.convert_local_pdf(
    "large_document.pdf",
    progress_callback=on_progress
)

远程 PDF 转换

from pdf_craft_sdk import PDFCraftClient, FormatType

client = PDFCraftClient(api_key="YOUR_API_KEY")

# Convert remote PDF to EPUB
download_url = client.convert(
    pdf_url="https://example.com/file.pdf",
    format_type=FormatType.EPUB,
    includes_footnotes=True
)
print(f"Download URL: {download_url}")

高级用法（分步执行）

from pdf_craft_sdk import PDFCraftClient, FormatType

client = PDFCraftClient(api_key="YOUR_API_KEY")

# Step 1: Upload file
cache_url = client.upload_file("document.pdf")

# Step 2: Submit conversion task
task_id = client.submit_conversion(
    cache_url,
    format_type=FormatType.MARKDOWN
)
print(f"Task submitted. ID: {task_id}")

# Step 3: Wait for completion
download_url = client.wait_for_completion(task_id)
print(f"Download URL: {download_url}")

配置选项

max_wait_ms: 最大等待时间（毫秒），默认 7200000（2小时）
check_interval_ms: 初始轮询间隔（毫秒），默认 1000
max_check_interval_ms: 最大轮询间隔（毫秒），默认 5000
backoff_factor: 轮询间隔递增倍数，或使用 PollingStrategy 枚举（默认 PollingStrategy.EXPONENTIAL / 1.5）
model: 用于转换的模型，默认 'gundam'

from pdf_craft_sdk import PDFCraftClient, FormatType, PollingStrategy

client = PDFCraftClient(api_key="YOUR_API_KEY")

# Example with custom polling configuration
download_url = client.convert(
    pdf_url="https://example.com/file.pdf",
    format_type=FormatType.MARKDOWN,
    model="gundam",
    includes_footnotes=True,
    ignore_pdf_errors=True,
    ignore_ocr_errors=True,
    max_wait_ms=600000,           # Maximum wait time: 10 minutes
    check_interval_ms=2000,       # Initial polling interval: 2 seconds
    max_check_interval_ms=10000,  # Maximum polling interval: 10 seconds
    backoff_factor=PollingStrategy.EXPONENTIAL  # or 1.5
)

TypeScript SDK

用于与 PDF Craft API 交互的 TypeScript SDK。它通过处理身份验证、任务提交和结果轮询，简化了将 PDF 转换为 Markdown 或 EPUB 的过程。

在 GitHub 上查看

安装

npm install pdf-craft-sdk-ts
# or
yarn add pdf-craft-sdk-ts

基本用法

import { PDFCraftClient, FormatType } from 'pdf-craft-sdk-ts';

const client = new PDFCraftClient("YOUR_API_KEY");

async function main() {
    try {
        const downloadUrl = await client.convert("cache://your-pdf-file.pdf", {
            formatType: FormatType.Markdown
        });
        console.log(`Success! Download: ${downloadUrl}`);
    } catch (error) {
        console.error("Error:", error);
    }
}

main();

本地文件上传与进度跟踪

import { PDFCraftClient, FormatType, UploadProgress } from 'pdf-craft-sdk-ts';

const client = new PDFCraftClient("YOUR_API_KEY");

const onProgress = (progress: UploadProgress) => {
    console.log(`Progress: ${progress.percentage.toFixed(2)}%`);
    console.log(`Part ${progress.currentPart}/${progress.totalParts}`);
};

// Upload and convert local file
const downloadUrl = await client.convertLocalPdf("document.pdf", {
    formatType: FormatType.Markdown,
    progressCallback: onProgress,
    includesFootnotes: true,
    uploadMaxRetries: 5
});

批量处理

import { PDFCraftClient, FormatType } from 'pdf-craft-sdk-ts';

const client = new PDFCraftClient("YOUR_API_KEY");

// Create batch conversion
const files = [
    { url: "cache://file1.pdf", fileName: "document1.pdf" },
    { url: "cache://file2.pdf", fileName: "document2.pdf" }
];

const batch = await client.createBatch(files, FormatType.Markdown, false);
const result = await client.startBatch(batch.batchId);

// Monitor progress
const status = await client.getBatchStatus(batch.batchId);
console.log(`Completed: ${status.completedCount}/${status.totalCount}`);

高级用法（分步执行）

import { PDFCraftClient, FormatType } from 'pdf-craft-sdk-ts';

const client = new PDFCraftClient("YOUR_API_KEY");

// Step 1: Submit task
const taskId = await client.submitConversion(
    "cache://file.pdf",
    FormatType.Markdown
);
console.log(`Task submitted. ID: ${taskId}`);

// Step 2: Poll for completion
const downloadUrl = await client.waitForCompletion(
    taskId,
    FormatType.Markdown
);
console.log(`Download URL: ${downloadUrl}`);

配置选项

formatType: FormatType.Markdown | FormatType.EPUB（默认: FormatType.Markdown）
model: 使用的模型（默认: 'gundam'）
wait: 是否等待完成（默认: true）
maxWaitMs: 最大等待时间（毫秒），默认 7200000（2小时）
checkIntervalMs: 初始轮询间隔（毫秒），默认 1000
maxCheckIntervalMs: 最大轮询间隔（毫秒），默认 5000
backoffFactor: 轮询间隔递增倍数，或使用 PollingStrategy 枚举（默认 PollingStrategy.Exponential / 1.5）

import { PDFCraftClient, FormatType, PollingStrategy } from 'pdf-craft-sdk-ts';

const client = new PDFCraftClient("YOUR_API_KEY");

// Example with custom polling configuration
const downloadUrl = await client.convert("cache://file.pdf", {
    formatType: FormatType.Markdown,
    model: "gundam",
    wait: true,
    includesFootnotes: true,
    ignorePdfErrors: true,
    ignoreOcrErrors: true,
    maxWaitMs: 600000,           // Maximum wait time: 10 minutes
    checkIntervalMs: 2000,       // Initial polling interval: 2 seconds
    maxCheckIntervalMs: 10000,   // Maximum polling interval: 10 seconds
    backoffFactor: PollingStrategy.Exponential  // or use a number like 1.5
});