PDF to MD conversion API

curl

# 1. Submit the conversion
curl -X POST https://hushvert.com/api/v1/conversions \
  -H "Authorization: Bearer $HUSHVERT_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{ "pair": "pdf-to-md", "bytes": '$(wc -c < input.pdf)' }'
# -> { "jobId": "...", "uploadUrl": "...", "pollUrl": "/api/v1/conversions/..." }

# 2. Upload the file to the presigned uploadUrl
curl -X PUT "$UPLOAD_URL" --data-binary @input.pdf

# 3. Poll until done, then download
curl https://hushvert.com/api/v1/conversions/$JOB_ID \
  -H "Authorization: Bearer $HUSHVERT_KEY"
# -> { "status": "done", "downloadUrl": "...", "expiresAt": "..." }
curl -o output.md "$DOWNLOAD_URL"

Node.js

import { readFile, writeFile } from 'node:fs/promises'

const KEY = process.env.HUSHVERT_KEY
const bytes = await readFile('input.pdf')

// 1. Submit
const submit = await fetch('https://hushvert.com/api/v1/conversions', {
  method: 'POST',
  headers: { authorization: `Bearer ${KEY}`, 'content-type': 'application/json' },
  body: JSON.stringify({ pair: 'pdf-to-md', bytes: bytes.byteLength }),
}).then((r) => r.json())

// 2. Upload to the presigned URL
await fetch(submit.uploadUrl, { method: 'PUT', body: bytes })

// 3. Poll until done
let job
do {
  await new Promise((r) => setTimeout(r, 1500))
  job = await fetch(`https://hushvert.com${submit.pollUrl}`, {
    headers: { authorization: `Bearer ${KEY}` },
  }).then((r) => r.json())
} while (job.status !== 'done' && job.status !== 'failed')

// 4. Download
const out = await fetch(job.downloadUrl).then((r) => r.arrayBuffer())
await writeFile('output.md', Buffer.from(out))

Python

import os, time, requests

KEY = os.environ["HUSHVERT_KEY"]
data = open("input.pdf", "rb").read()

# 1. Submit
submit = requests.post("https://hushvert.com/api/v1/conversions",
    headers={"Authorization": f"Bearer {KEY}"},
    json={"pair": "pdf-to-md", "bytes": len(data)}).json()

# 2. Upload to the presigned URL
requests.put(submit["uploadUrl"], data=data)

# 3. Poll until done
job = {}
while job.get("status") not in ("done", "failed"):
    time.sleep(1.5)
    job = requests.get(f"https://hushvert.com{submit['pollUrl']}",
        headers={"Authorization": f"Bearer {KEY}"}).json()

# 4. Download
out = requests.get(job["downloadUrl"]).content
open("output.md", "wb").write(out)

How do I convert PDF to MD with an API?

Send a POST to /api/v1/conversions with the pair and the byte count, upload the file to the presigned URL it returns, then poll until the job is done and download the result. It is three short requests; full code is above in curl, Node, and Python.

Why not just convert PDF to MD myself?

You can, if you want to run and maintain the right native conversion toolchain. This conversion needs a server-side toolchain a browser cannot run. The API gives it to you as a single HTTP call, with the infrastructure managed for you.

What does it cost?

Every account gets 50 free conversions per month across all of its keys. Beyond that, the API draws on account credits, billed per conversion. Everything a browser can do is free and keyless in the open-source SDK.

What happens to my file?

The file uploads over an encrypted connection. The input is deleted the moment the conversion finishes; the converted output is kept for about an hour so you can download it, then deleted too.

Can my AI agent do this conversion?

Yes. The @hushvert/mcp Model Context Protocol server gives Claude Code, Cursor, and other agents a convert_file tool over this same API, so an agent can run the conversion in one call.