Support & known limitations
LessTokens is a free beta. Here is how to reach us, what the tool cannot do yet, and where to tell us what you need.
Contact support
For questions, problems, privacy requests, or takedown reports, email us:
Known limitations
- Screenshots, images, and scanned PDFs are all read with self-hosted OCR. Scanned PDFs are read for the first 20 pages, and anything beyond that is disclosed in the output; accuracy depends on scan quality and text clarity, and the result carries the engine's confidence in the characters it read (not how much of the page it found). When OCR finds no usable text the tool declines honestly instead of returning empty output. Vertical (top-to-bottom) Chinese recognition quality is not yet validated.
- Always review the generated Markdown before relying on it. Conversion is automated and can miss or misread content, especially complex layouts.
- Free beta limits apply: up to 100 MB per file, up to 500 PDF pages, and a daily conversion cap.
- Complex tables are kept as readable text rather than guessed — some table structure may be simplified.
- Token counts are estimates (GPT-family tokenizer approximation) and vary by AI model. A reduction is not guaranteed for every file.
Request a feature or send feedback
Missing something? Tell us what you need — it directly shapes what we build next.
Want more? Tell us what to build next
These features are not available yet. Pick what you need and leave your email — we will let you know when it ships. Optional, and never required to use the converter.