The API behind PDFTrusted

PDFTrusted does not sell a developer API today. What it does have is a working HTTP surface that the whole product runs on: quota checks, upload presigning, conversion jobs, OCR, share links and embeds. Those routes are documented here because hiding them does not make them safer, and because an honest reference is more useful than a waitlist.

Two things follow from that. First, every example below was captured from the real service and pasted verbatim, apart from one redaction noted in the example itself. Second, there is no public key programme, so nothing here is a promise of stable access for third parties — if you need that, the contact link at the top is the honest route.

The endpoints are session-scoped: a request carries the same browser session the app uses, whether that is an anonymous visitor or a signed-in account. Plan entitlements (Free or Pro) are read from that account, so Pro limits follow the session rather than a separate key.

• No API key exists today, and no endpoint accepts one
• Quota and rate limits are enforced per device and per job, not per key
• Account endpoints (history, export, deletion) require a signed-in session
• Billing webhooks are signed by the payment provider, not by this API

Server conversions are three calls. Presign, upload, enqueue — then poll or wait. These are the real shapes, with the signed URL shortened because it is single-use and expires in five minutes.

Step 1 — ask for an upload slot

curl -X POST https://pdftrusted.com/api/v1/convert/presign \
  -H 'Content-Type: application/json' \
  -d '{"tool":"pdf-to-word","filename":"contract.pdf",
       "contentType":"application/pdf","sizeBytes":1024}'

{
  "uploadId": "0e02fadd-8268-4d90-b3bd-e0443cd33e35",
  "uploadUrl": "https://<single-use-upload-url>?X-Amz-Expires=300&…",
  "contentType": "application/pdf",
  "jobId": "163ba447-a06a-4091-99c9-6aed3c3894cd"
}

Step 2 — PUT the bytes to that URL

curl -X PUT "<uploadUrl from step 1>" \
  -H 'Content-Type: application/pdf' \
  --data-binary @contract.pdf

Step 3 — enqueue, then read the result

curl -X POST https://pdftrusted.com/api/v1/convert/enqueue \
  -H 'Content-Type: application/json' \
  -d '{"uploadId":"0e02fadd-…","tool":"pdf-to-word"}'

A finished job returns the download link and timing rather than the file itself, so the browser can fetch the result directly:

{
  "downloadUrl": "https://<signed-download-url>",
  "fileName": "contract.docx",
  "mimeType": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
  "jobId": "163ba447-a06a-4091-99c9-6aed3c3894cd",
  "durationMs": 2841
}

Every conversion response also carries X-Convert-Job-Id and X-Convert-Duration-Ms headers, and is marked Cache-Control: no-store.

Browser-side tools check quota first. This response is verbatim from a Free session on October 18, 2026:

{
  "ok": true,
  "jobsRemaining": 20,
  "pagesRemaining": 500,
  "bytesRemaining": 524288000,
  "nearLimit": false
}

Exceed a ceiling and the same endpoint answers 413 with a machine-readable code instead of prose alone:

{
  "error": "This file exceeds the Free limit of 100 MB.",
  "code": "FREE_FILE_LIMIT"
}

These come from the same constants the product enforces, so they cannot drift away from what the interface shows.

Per-tool page ceilings sit underneath these: merge allows 75 pages, compression 50, rotation, cropping and split-by-size 150. A file over a per-tool ceiling is rejected before any work starts.

• No public API keys, sandbox keys or self-serve signup for server access
• No SLA and no published uptime figure — the status page shows what is up now, not a percentage we cannot prove
• No batch endpoint and no conversion webhooks; polling a job is how you get a result
• No separate API pricing, because there is no separate API product yet
• No third-party processing credits: the free and Pro quotas in the table above are the whole story

Smallpdf’s footer links its “Developers” entry to a page listing the third-party libraries powering its converters, not to an API reference. iLovePDF sells a separate API product with its own pricing and docs. We are doing neither: no key shop, no vendor list — just the routes this product already runs, documented with the responses they actually return.

Comparisons of the tools themselves, with prices and sources, are on the comparison pages.