Fit a long document into your AI's context window
Paste a long report or contract into ChatGPT or Claude and it often gets cut off, costs a pile of tokens, or the AI quietly skips the middle. Clean the document first — and optionally shrink it — so more of what matters actually reaches the model.
Why long documents break down in AI
Length itself isn't the only problem — most of what you paste isn't even content:
- The document is longer than the context window, so the tail is silently dropped and the AI answers from a partial read.
- Raw exports are padded with page headers, footers, page numbers, and repeated boilerplate that burn tokens without adding meaning.
- On paid APIs every extra token costs money, so a bloated document gets expensive fast.
- Even when it fits, content buried in the middle gets less of the model's attention than the start and end.
The fix: strip the noise, then optionally shrink
LessTokens first removes the layout noise — page furniture, decorative rules, broken line breaks — losslessly, keeping every word of your real content. If you need it smaller still, an opt-in Lean mode also drops repeated boilerplate and watermarks. Either way you get clean Markdown and an estimated token reduction, so you can see how much lighter it got before you paste.
Drop a document — or paste a screenshot
PDF, Word, PowerPoint, Excel, HTML, CSV/TSV, TXT, or a screenshot (PNG/JPG — paste with Ctrl/⌘+V) · up to 20 MB
Have one specific format? Use the universal converter →
What LessTokens does to help it fit
- Strips repeated page headers, footers, and standalone page numbers that add tokens without meaning.
- Removes decorative separators and fixes broken line breaks, so the text is compact and clean.
- Optional Lean mode removes repeated boilerplate and watermarks across the whole document — opt-in, and it lists what it took out.
- Shows an estimated token reduction, and in Standard mode never summarizes or invents — every word of your real content stays.
How much smaller — honestly
For already-clean prose the reduction is small: clean text is close to incompressible without deleting meaning, and we will not do that in Standard mode. The real wins are on documents heavy with repeated furniture, watermarks, or boilerplate — that is what Lean mode targets, and it removes repeated content verbatim, never rewording or summarizing. If a document is far larger than the context window, splitting it into parts is sometimes still necessary; LessTokens never silently drops your unique content to force a fit.
Questions
Will this let me fit any document into the context window?
It removes the tokens that don't carry meaning — page furniture and, in Lean mode, repeated boilerplate — so more of your real content fits. But if the actual content is still larger than the window, no cleanup can shrink it without deleting meaning; splitting it into parts is the honest option, and we never do that silently.
Does Lean mode change my words?
No. Lean mode only removes repeated content (headers, watermarks, boilerplate) verbatim — it never rewords, paraphrases, or summarizes. It also lists what it removed so you can verify. Standard mode keeps every line.
Is my file private?
Files are processed temporarily, never used to train AI, and automatically deleted within 60 minutes. See the privacy policy for the full detail.
File retention
Uploaded originals and conversion results are temporary. They are not kept as a permanent document library.
- Uploaded originals are used only to produce your Markdown result and are deleted as soon as processing ends, or within 60 minutes at the latest.
- Conversion results and job records are available for up to 60 minutes, then expire and are automatically removed from the active database.
- We do not keep a permanent document library. Limited provider-managed infrastructure snapshots may retain residual copies for up to 5 days before automatic expiry.
- Uploaded originals are held in temporary local storage only while the job is processed. Recurring cleanup removes any expired temporary upload left by an interruption.