About

PDF Conversion API is an HTTP API that converts files. You send a file to POST /v2/convert/{from}/to/{to} as multipart/form-data, and the converted file comes back in the JSON response, as base64 or, with StoreFile=true, as a download link. The same request shape covers 556 conversions: Word, Excel and PowerPoint files, PDFs, images including iPhone HEIC photos, CAD drawings, email, e-books and web pages. The conversions list has every pair.

Who runs it

Zeel Shekhaliya, Founder. Zeel Shekhaliya is the founder of PDF Conversion API and runs it from Surat, Gujarat, India. Zeel signs the conversion guides and test reports on this site; every number in them comes from a run of the API that you can repeat with the published sample files.

Customers contract with Zeel Shekhaliya, an individual based in Surat, Gujarat, India, not with a company. The contact page lists the mailboxes we read and what each one is for.

Where things stand

Anyone can get a free API key by signing in with Google or GitHub at app.pdfconversionapi.com. The Free plan allows 250 conversions a month, files up to 10 MB and 1 conversion at a time. Billing is not open yet, so the Free plan is the only one you can start yourself. For higher limits before billing opens, write to sales@pdfconversionapi.com.

These are not built yet:

• Billing. The pricing page shows the prices we plan to charge when billing opens.
• A dashboard. The key page shows your key once, lets you replace it, shows this month's usage and is where you delete your account. There is nothing else yet.
• Team accounts. Each account belongs to one verified email address.
• A status page.
• Client libraries (SDKs). Any HTTP client that can send multipart/form-data works.
• Async jobs and webhooks. Every request is answered in the same HTTP call.
• Input from a URL. You upload the file itself.
• Request bodies other than multipart/form-data, such as JSON with base64-encoded files.

What it is for

Converting office files on your own servers means installing an office suite such as Microsoft Office, keeping fonts that match the documents, stopping conversions that hang, and guarding against files made to crash or attack the converter. The API does that work behind one request. These are the rules we hold it to:

• If you change an option that ConvertAPI documents for a conversion and our converter doesn't support it, you get a 400 that names the option instead of a result that ignores it.
• We don't keep your files unless you ask. With StoreFile=true the download link works for 3 hours, and a cleanup job that runs every 15 minutes then deletes the file. The security page has the details.
• Each conversion page says what the conversion loses, and its sample output and timing come from running our API, not from an estimate. How we test explains the scripts.
• When something isn't built, we say "not yet".

If your code already posts multipart forms to ConvertAPI-style endpoints, it can call our API after you change the URL and the key, as long as it uses conversions and options we support: 279 of the 328 endpoints ConvertAPI documents work here with the same {from}/to/{to} path and do the same job. ConvertAPI's own client libraries upload files in a separate step that we don't support, so they don't work with our API yet. The comparison page lists what matches and what doesn't.

Where it runs

The API and this website run on one cloud provider. The subprocessor list names it and shows what each of its services handles.

• Gateway. A gateway on the provider's network checks your key against SHA-256 hashes kept in the provider's key-value store, applies the limit of 600 requests a minute per account and your plan's limits (conversions a month, upload size and conversions at once), and sends each conversion to a container pool that has the right converter.
• Conversions. Three pools of conversion containers, each with a different set of converters. The containers run as a non-root user with internet access turned off, and they accept only requests that carry the gateway's own key.
• Stored results. Files you ask us to keep with StoreFile=true go to the provider's object storage, in its US jurisdiction.
• Usage records. Each conversion request the gateway passes to a pool writes one usage record to the provider's analytics store: account, plan, formats, pool, status code, duration and upload size. File contents and file names are never part of it.
• Account pages. The pages at app.pdfconversionapi.com are a separate small app at the same provider. When you sign in with Google or GitHub, we keep only your verified email address and that account's id. The app keeps those, your sign-in sessions, and your key's SHA-256 hash and last four characters in a database at the provider.
• This website. Static pages served by the same provider.

Processing isn't pinned to a region. A conversion runs in whichever of the provider's locations has a container ready, which can be outside the US.

How conversions are handled

• Converters that run as separate programs start a new process for every conversion, in a new temporary folder that is deleted afterwards.
• Some of the converters are described by their own publishers as alpha or early-development software, so each conversion page lists what we saw the conversion lose.
• When we build our image, each converter that we take as a release archive is checked against the SHA-256 checksums its project publishes, and a mismatch stops the build.
• Text recognition (OCR) reads English only. Its two neural-network models came pretrained, and our build pins both model files by SHA-256 and won't build with any other file. The models read your file to find text, and nothing you send is used to train them.

ConvertAPI is a trademark of its owner. pdfconversionapi.com is not affiliated with it.