RTF to HTML Cloud Function

RTF in, clean HTML out: a Java 21 Google Cloud Function that converts uploaded RTF documents into self-contained HTML, powered by the ScroogeXHTML RTF converter library.

Features

• Rich formatting — fonts, paragraph indents and borders, tables with cell borders, hyperlinks and bookmarks.
• Self-contained output — JPG, PNG and BMP pictures up to 256 KB are embedded as Base64 data URIs.
• Validated output — each document is checked locally with the Nu Html Checker; a footer shows the converter version, the conversion time and either the HTML5 logo or the number of validation errors.
• Safe by default — link targets are allow-list filtered by the converter, every response carries a restrictive Content-Security-Policy, and the demo pages show the result in a sandboxed iframe.
• Bounded resources — uploads are limited to 1 MB and read with a hard cap.
• Origin-restricted — only requests from https://www.scroogexhtml.com are accepted (CORS).

Requirements

• JDK 21+ and Maven 3.8+
• Google Cloud CLI (gcloud) for deployment
• ScroogeXHTML 11.5.0 and ScroogeXHTML-Pictures 11.5.0 in your Maven repository — see scroogexhtml.com

Quick start

Build the shaded JAR:

mvn clean package

Run the function locally on http://localhost:8080/:

mvn function:run

Convert a document:

curl -H "Origin: https://www.scroogexhtml.com" -F "file=@document.rtf" http://localhost:8080/

Note

Browsers send their own Origin header, which the function rejects unless it matches ALLOWED_ORIGIN in RtfToHtml.java. To try demo-local.html in a browser, serve it from a local web server and set ALLOWED_ORIGIN to that server's origin.

API

POST / with a multipart/form-data body. The first part whose file name ends in .rtf is converted; the form field name does not matter.

Status	Meaning
200	Converted document (text/html; charset=utf-8)
204	CORS preflight (OPTIONS) from the allowed origin
400	Not multipart/form-data, malformed multipart body, or no .rtf file
403	Missing or disallowed Origin
405	Method other than POST or OPTIONS
413	File larger than 1 MB

Errors are returned as text/plain messages.

Deployment

Deploy the contents of target/ as a 2nd gen function, or run deploy.cmd on Windows:

gcloud functions deploy rtf-to-html-demo-11 \
  --gen2 \
  --runtime=java21 \
  --region=europe-west10 \
  --entry-point=demo.RtfToHtml \
  --source=./target \
  --trigger-http \
  --allow-unauthenticated \
  --max-instances=3 \
  --cpu=1 \
  --memory=512Mi

demo-prod.html calls the deployed endpoint.

One full vCPU keeps the response time reasonable: on the default instance size (a fraction of a vCPU), a 35 KB document took about 3.5 s, mostly for HTML validation. Files below 128 KB are converted five times, and the footer shows the median conversion time.

Files over 1 MB are rejected before conversion. A file close to that limit can still take longer to convert than the default 60 s request timeout (seen with 1 MB of unformatted text on the default instance size); increase --timeout if you need more headroom.

Configuration

All settings are constants and converter options in RtfToHtml.java:

Setting	Default
ALLOWED_ORIGIN	https://www.scroogexhtml.com
MAX_SIZE_BYTES	1 MB
CONTENT_SECURITY_POLICY	default-src 'none', inline styles and data: images allowed
Embedded picture limit	256 KB (MemoryPictureAdapterDataURI)
Table cell borders	Taken from the RTF document (setConvertCellStyles)

The converter options (tables, hyperlinks, CSS, font size unit, …) are documented at scroogexhtml.com.

Learn more

• ScroogeXHTML RTF converter — features, documentation and downloads
• RTF to HTML in the cloud — background article on this project
• Google Cloud Functions for Java

License

Mozilla Public License 2.0