Home › Catalog › Diff Tool › Why Forge Diff
Diffchecker.com is the household name for online diffs and the free tier covers casual use. The Forge Diff API is what you reach for when you want to script it, diff a large file without a save-and-share paywall, or get structured JSON opcodes that a downstream tool can consume.
| Feature | Forge Diff | Diffchecker (free tier) |
|---|---|---|
| Text diff (multiple formats) | ✓ unified, context, ndiff, side-by-side HTML, JSON opcodes | ✓ side-by-side, inline |
| JSON canonical diff | ✓ sorted-keys, structured summary | ✓ (separate tool) |
| REST API for scripts & CI | ✓ free, no key | × (paid Pro plan only) |
| Max input size | 200,000 chars per side | limited; longer inputs pushed to paid tier |
| Ignore whitespace / ignore case toggles | ✓ | ✓ |
| Ads / interstitials | × | ✓ on free tier |
| Signup required | × | × for casual use, ✓ for save/share |
| Cost | $0 | $0 free / paid Pro for API, image, folder, PDF diff |
Diffchecker has a Pro API behind a paid subscription. Forge Diff is a free public endpoint. Compute a unified diff from anywhere:
curl -X POST https://forge-node.tail2b516d.ts.net/diff/text \
-H 'Content-Type: application/json' \
-d '{"a":"hello world","b":"hello forge","format":"unified"}'
Five output formats: unified for the patch shape git already speaks, context for context-diff style, ndiff for character-level annotation, html for a colored side-by-side surface, and json for a structured opcode list a downstream tool can consume.
Comparing two JSON blobs and getting a meaningful answer is harder than text diff. Forge Diff canonicalizes both sides (sorted keys, normalized whitespace, stable number formatting) before diffing, then returns both the line-level diff and a structured summary of added / removed / changed keys. Drop in two API responses, get a stable diff regardless of key ordering.
curl -X POST https://forge-node.tail2b516d.ts.net/diff/json \
-H 'Content-Type: application/json' \
-d '{"a":{"x":1,"y":2},"b":{"y":2,"x":3}}'
The format=html response returns Python's built-in difflib.HtmlDiff table — clean, lightweight, and embeddable as an iframe or with a single fetch. No tracking pixel, no analytics call, no banner.
Need a single number, not a diff? POST /diff/similarity returns a 0.0-1.0 ratio using SequenceMatcher. Useful for plagiarism heuristics, fuzzy match thresholds, or test-output approximate comparisons.
curl; fail the build when the unified diff is non-empty.Yes. No signup, no API key. Rate-limited per IP at 60 requests per minute; documented in the Diff API docs.
200,000 characters per side. That covers any single source file and most small documents. For larger diffs, split client-side and stream.
No. Diffs are computed in-process and the response is returned synchronously. Standard nginx-style access logs (IP, path, status) only. No diff body is persisted, which is also why we don't expose a save-and-share URL.
Not yet. Those are Diffchecker Pro features; we are focused on the text and structured-data cases that scripts and pipelines actually want.
Try the Diff tool, or browse the full free API catalog for 15 more developer APIs.