Prompt Diff — For Fine-Tuning Datasets

Compare two prompts side-by-side with word-level diff highlighting. Shows added/removed words plus the token count delta — useful when iterating on a prompt and you need to confirm a change shrank or grew the token budget. Built on the same Myers algorithm git uses.

How to use the Prompt Diff — For Fine-Tuning Datasets

Paste two prompt versions. The diff highlights word-level changes and reports the character / estimated-token delta. For exact token counts, run each version through the OpenAI Token Counter separately.

Seeing what changed between two prompts

Prompt engineering is iterative: you tweak a wording, add a constraint, drop a sentence, and rerun. After a few rounds it's hard to remember exactly what changed between the version that worked and the one that didn't — and small edits can quietly push the prompt over a token budget.

This diff compares two prompt versions with word-level highlighting, marking what was added and removed, and reports the character and estimated-token delta so you can see at a glance whether a change grew or shrank the prompt. It uses the same Myers diff algorithm git relies on. For an exact token count of either version, run it through the OpenAI token counter.

Common use cases

  • Prompt iteration — confirm exactly which words changed between two attempts.
  • Token budgeting — check whether an edit grew or shrank the prompt size.
  • Reviewing edits — see a teammate's prompt change as a clear added and removed diff.
  • Regression hunting — compare a working version against one that started failing.
  • Documenting changes — capture what a prompt revision actually altered.

Frequently asked questions

What does the diff highlight?
Word-level additions and removals between the two versions, plus the character count and an estimated-token delta so you can gauge the size impact of the change.

How are tokens estimated?
From the character count using a typical ratio — close enough to judge a change. For an exact figure, run each version through the OpenAI token counter.

What algorithm does it use?
The Myers diff algorithm, the same one git uses, which produces a minimal, readable set of additions and deletions.

Is it line-based or word-based?
Word-based, which suits prose prompts better than line diffs — it highlights the specific words that changed rather than whole lines.
Embed this tool on your site

Free to embed, no attribution required (but appreciated). Paste this where you want the tool to appear: