Turn Markdown ZIP packages into production-ready HTML with async import
Upload one ZIP archive with Markdown and local images, create an async import job, and get back HTML once processing finishes. The system handles assets, fixes references, and prepares output for preview, distribution, and downstream workflows.
Import an article and its images together
- ZIP import without manually uploading every image
- Images are processed and uploaded, while the ZIP itself never goes to COS
- Directory paths, theme styles, saved articles, and AI-readable docs in one flow
Async ZIP ingestion
Designed for Obsidian and Markdown package exports with relative image references and queue-based processing.
Final output after job completion
Create a job, poll its status, and receive HTML you can preview, publish, or hand off downstream without extra manual cleanup.
Theme-aware rendering
Keep output styling consistent across brand content, knowledge base publishing, and batch workflows.
Auto-organized assets
Articles and assets stay neatly organized so search, reuse, and team collaboration remain manageable.
Typical use cases
- 1
Package Markdown articles from Obsidian, a knowledge base, or local editorial files into a ZIP archive and upload it first.
- 2
Create an async import job so the system can handle image uploads, reference replacement, and layout conversion automatically.
- 3
Poll until the job finishes, then feed the generated result into publishing, distribution, AI enrichment, or internal editing pipelines.
What you get in practice
- HTML content returned after job completion, ready for preview, embedding, or downstream publishing
- Article material with image references already cleaned up for further editing
- A more stable asset structure for collaboration, archiving, and batch management
- Consistent visual output across Markdown packages from different sources
API keys and AI-readable docs
The section below works as both a developer console and a copy-ready async integration handoff for AI or engineering teams.
ZIP import API docs
Create an API key in one click, then copy the Markdown guide below for your AI agent or engineering teammate to implement the async workflow.
Create import API key
Please login to access this feature
Existing keys
No API keys yet
Theme API identifiers
Each themeId matches the theme name shown in the editor. Human-readable names and reference colors are listed below.
AI-readable API integration guide
Imgto.link Markdown ZIP Import API
What this API does
- Upload one ZIP package containing one Markdown file and its local images
- Queue an asynchronous import job
- Automatically create the target image directory if it does not exist
- Automatically save the imported Markdown into the article folder
公众号 - Upload referenced images to Imgto.link storage
- Replace local image references in Markdown
- Return the final resolved Markdown, HTML output, saved article info, and image directory path after the job completes
Base URL
https://api.imgto.link
Endpoints on the same host
- ZIP upload endpoint:
POST https://api.imgto.link/uploads/md-import - Create import job:
POST https://api.imgto.link/api/v1/md/import - Poll job status:
GET https://api.imgto.link/api/v1/md/import/jobs/{jobId} - In production you can proxy the control-plane routes on the same
https://api.imgto.linkhost, so clients only need one public base URL.
Authentication
Use an API key in the Authorization header:
Authorization: Bearer YOUR_API_KEY
Async workflow
- Upload the ZIP archive to the upload endpoint and get back
uploadRelativePath - Create an async import job with
POST https://api.imgto.link/api/v1/md/import - Poll
GET https://api.imgto.link/api/v1/md/import/jobs/{jobId}until the job reaches a terminal status
Step 1: upload ZIP archive
Endpoint
POST https://api.imgto.link/uploads/md-import
Content-Type
multipart/form-data
Form fields
archive(required): ZIP file containing one Markdown file and its local image files
Success response
{
"code": 200,
"data": {
"relativePath": "inbox/2026/03/12/uuid-article-package.zip",
"originalFilename": "article-package.zip",
"size": 345678,
"uploadedAt": "2026-03-12T08:00:00.000Z"
}
}
Step 2: create import job
Endpoint
POST https://api.imgto.link/api/v1/md/import
Content-Type
application/json
Request body
{
"uploadRelativePath": "inbox/2026/03/12/uuid-article-package.zip",
"archiveFilename": "article-package.zip",
"archiveSize": 345678,
"themeId": "classic",
"directoryPath": "public/articles/demo",
"entryMarkdown": "docs/index.md",
"saveFolderId": 12,
"mdfileId": 123
}
Request field notes
uploadRelativePath(required): value returned from step 1archiveFilename(required): original zip filenamearchiveSize(optional): archive size in bytesthemeId(optional): theme API name, default isclassicdirectoryPath(optional): base upload directory path likepublic/articles/launch-2026entryMarkdown(optional): specific markdown file path inside the ZIP when there are multiple markdown filessaveFolderId(optional): target markdown folder ID in the editormdfileId(optional): existing markdown article ID. When provided, the import is treated as a replace/update. The old article and its imported images are deleted before the new content is saved
Success response
{
"code": 202,
"message": "Accepted",
"data": {
"jobId": "mdi_1234567890abcdef",
"status": "QUEUED",
"progress": 0,
"pollUrl": "/api/v1/md/import/jobs/mdi_1234567890abcdef"
}
}
Step 3: poll job status
Endpoint
GET https://api.imgto.link/api/v1/md/import/jobs/{jobId}
Non-terminal response example
{
"code": 200,
"message": "Ok",
"data": {
"jobId": "mdi_1234567890abcdef",
"status": "PROCESSING",
"progress": 65,
"currentStage": "processing_images",
"message": "正在处理图片 2/4",
"compensationStatus": "NOT_REQUIRED",
"updatedAt": "2026-03-12T08:01:15.000Z"
}
}
Completed response example
{
"code": 200,
"message": "Ok",
"data": {
"jobId": "mdi_1234567890abcdef",
"status": "SUCCEEDED",
"progress": 100,
"currentStage": "completed",
"message": "导入完成",
"compensationStatus": "NOT_REQUIRED",
"updatedAt": "2026-03-12T08:01:30.000Z",
"result": {
"mdfileId": 123,
"resolvedMarkdown": "# Title\n\n",
"html": "<section>...</section>",
"wechatApiHtml": "<section>...</section>",
"imageDirectoryPath": "public/公众号/my-article",
"savedArticle": {
"id": 123,
"mdfileId": 123,
"title": "My Article",
"folder": {
"id": 9,
"name": "公众号",
"fullPath": "公众号"
}
}
}
}
}
Failed response example
{
"code": 200,
"message": "Ok",
"data": {
"jobId": "mdi_1234567890abcdef",
"status": "FAILED_COMPENSATED",
"progress": 100,
"currentStage": "failed",
"message": "导入失败,已回滚上传图片与用户额度",
"compensationStatus": "COMPLETED",
"updatedAt": "2026-03-12T08:02:00.000Z",
"error": {
"message": "Referenced image not found: ./images/cover.png"
}
}
}
Job status notes
- Terminal success status:
SUCCEEDED - Terminal failure status:
FAILED_COMPENSATED - Transitional failure statuses can appear briefly during rollback, such as
FAILED_PENDING_COMPENSATIONandFAILED_COMPENSATION_ERROR resultis returned only when the job succeedserroris returned when the job fails
Result field notes
result.resolvedMarkdown: Markdown after local image paths are replaced with Imgto.link URLsresult.html: Base rendered HTML for preview or debuggingresult.wechatApiHtml: Recommended HTML for submitting to the WeChat Official Account APIcontentfieldresult.mdfileId: The saved Markdown article ID, returned for convenient reuse in the next replace import requestresult.savedArticle: Saved Markdown article record in the editorresult.savedArticle.mdfileId: Stable article identifier to pass back on the next import if you want to replace this article instead of creating a new oneresult.imageDirectoryPath: Final image upload directory path
Example cURL: step 1 upload
curl -X POST "https://api.imgto.link/uploads/md-import" \
-F "archive=@article-package.zip" \
Example cURL: step 2 create job
curl -X POST "https://api.imgto.link/api/v1/md/import" \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"uploadRelativePath": "inbox/2026/03/12/uuid-article-package.zip",
"archiveFilename": "article-package.zip",
"archiveSize": 345678,
"themeId": "classic",
"directoryPath": "public/articles/demo"
}'
Example cURL: step 3 poll job
curl -X GET "https://api.imgto.link/api/v1/md/import/jobs/mdi_1234567890abcdef" \
-H "Authorization: Bearer YOUR_API_KEY"
Example JavaScript: end-to-end async import
const uploadFormData = new FormData();
uploadFormData.append("archive", zipFile);
const uploadResponse = await fetch("https://api.imgto.link/uploads/md-import", {
method: "POST",
body: uploadFormData
});
const uploadResult = await uploadResponse.json();
const relativePath = uploadResult.data.relativePath;
const createResponse = await fetch("https://api.imgto.link/api/v1/md/import", {
method: "POST",
headers: {
Authorization: "Bearer YOUR_API_KEY",
"Content-Type": "application/json"
},
body: JSON.stringify({
uploadRelativePath: relativePath,
archiveFilename: uploadResult.data.originalFilename,
archiveSize: uploadResult.data.size,
themeId: "classic",
directoryPath: "public/articles/demo"
})
});
const createResult = await createResponse.json();
const jobId = createResult.data.jobId;
let finalResult = null;
for (;;) {
const pollResponse = await fetch("https://api.imgto.link/api/v1/md/import/jobs/" + jobId, {
method: "GET",
headers: {
Authorization: "Bearer YOUR_API_KEY"
}
});
const pollResult = await pollResponse.json();
const job = pollResult.data;
if (job.status === "SUCCEEDED") {
finalResult = job.result;
break;
}
if (job.status === "FAILED_COMPENSATED") {
throw new Error(job.error?.message || "Import failed");
}
await new Promise((resolve) => setTimeout(resolve, 1800));
}
console.log(finalResult.mdfileId);
console.log(finalResult.resolvedMarkdown);
console.log(finalResult.wechatApiHtml);
console.log(finalResult.savedArticle);
Example: replace an existing imported article
Use the result.savedArticle.mdfileId returned by the previous successful import job.
const replaceResponse = await fetch("https://api.imgto.link/api/v1/md/import", {
method: "POST",
headers: {
Authorization: "Bearer YOUR_API_KEY",
"Content-Type": "application/json"
},
body: JSON.stringify({
uploadRelativePath: uploadedZip.data.relativePath,
archiveFilename: uploadedZip.data.originalFilename,
archiveSize: uploadedZip.data.size,
themeId: "classic",
mdfileId: previousResult.savedArticle.mdfileId
})
});
const replaceResult = await replaceResponse.json();
console.log(replaceResult.data.jobId);
Replace behavior
mdfileIdomitted: create a new import job using the uploaded ZIPmdfileIdprovided: force this import to replace that exact article- During replacement, the job deletes the old markdown record first
- It also deletes images previously imported for that markdown article before uploading the new images
- Read the latest
mdfileIdfrom the final successful poll result and keep using it for future updates
Example: publish to WeChat Official Account draft API
Use result.wechatApiHtml from the completed import job for the content field instead of result.html.
const completedImport = await waitForImportJob(jobId);
const wechatApiHtml = completedImport.result.wechatApiHtml;
const draftPayload = {
articles: [
{
title: completedImport.result.savedArticle?.title || "Imported article",
author: "Your Team",
digest: "",
content: wechatApiHtml,
content_source_url: "",
thumb_media_id: "YOUR_WECHAT_THUMB_MEDIA_ID",
need_open_comment: 0,
only_fans_can_comment: 0
}
]
};
const wechatResponse = await fetch(
"https://api.weixin.qq.com/cgi-bin/draft/add?access_token=YOUR_WECHAT_ACCESS_TOKEN",
{
method: "POST",
headers: {
"Content-Type": "application/json"
},
body: JSON.stringify(draftPayload)
}
);
const wechatResult = await wechatResponse.json();
console.log(wechatResult);
WeChat publishing note
- Prefer
result.wechatApiHtmlfor WeChat Official Account API publishing - Do not use
result.htmlas thecontentfield if you need output aligned with the website's copy/export flow - Image URLs inside
result.wechatApiHtmlshould already point to uploaded Imgto.link assets - You still need to provide your own WeChat
access_tokenandthumb_media_id
Theme API names
- Default Theme (
warm): #ea580c, #333 - Chinese Style (
china): #7f1d1d, #292524, #f5f5f4 - Cyberpunk (
cyber): #ec4899, #a855f7, #3b82f6 - WeChat Classic (
classic): #07c160, #333 - Ink Serif (
ink): #3f3f46, #78716c - Editorial Magazine (
editorial): #1f2937, #ef4444 - Healing WenKai (
wenkai): #57534e, #78716c - Deep Reading (
business): #1e3a8a, #1e40af - Forest Trail (
forest): #059669, #ecfdf5 - Azure Coast (
azure): #0ea5e9, #e0f2fe - Cozy Coffee (
coffee): #78350f, #fffbeb - Cyber Neon (
neon): #d946ef, #06b6d4 - Oriental Red (
red): #be123c, #fff1f2 - Silver Tech (
silver): #475569, #e2e8f0 - Minimal Blanc (
minimal): #000, #999 - Morandi Pink (
morandi): #be185d, #fbcfe8 - Sunny Vibes (
sunny): #f97316, #fbbf24 - Mint Summer (
mint): #10b981, #d1fae5 - Golden Autumn (
autumn): #d97706, #fef3c7 - Provence (
lavender): #8b5cf6, #ede9fe - Vintage Film (
vintage): #7f1d1d, #fecaca - Study Notes (
notes): #facc15, #fef08a - Smoky Grey (
grey): #374151, #f3f4f6
ZIP package rules
- Include exactly one Markdown file, or pass
entryMarkdown - Reference local images with relative paths such as
./images/a.png - Remote image URLs are kept as-is
- The API converts static images to AVIF by default
- The imported Markdown is automatically saved to the article folder
公众号 - The image upload directory becomes
{directoryPath}/{article-slug}
AI integration instruction
If you give this API to an AI coding assistant, tell it:
"Use the async workflow. First upload the ZIP to https://api.imgto.link/uploads/md-import with multipart/form-data and read data.relativePath, data.originalFilename, and data.size. Then call POST https://api.imgto.link/api/v1/md/import with JSON using uploadRelativePath, archiveFilename, and optional themeId, directoryPath, entryMarkdown, saveFolderId, and mdfileId. Authenticate the control-plane API with Authorization: Bearer YOUR_API_KEY. Then poll GET https://api.imgto.link/api/v1/md/import/jobs/{jobId} until status becomes SUCCEEDED or FAILED_COMPENSATED. On success, read result.resolvedMarkdown, result.wechatApiHtml, result.savedArticle, result.mdfileId, and result.imageDirectoryPath. Prefer result.wechatApiHtml when publishing to the WeChat Official Account API content field."