Token Span Label Workbench
Convert supplied Unicode-scalar entity spans and token boundaries to strict IOB2/BILUO labels or back.
Loading tool…
How to use Token Span Label Workbench
Inspect an annotation exchange with explicit source coordinates instead of relying on hidden tokenization. The workbench validates transitions, nonoverlap and exact token-edge alignment, then exports both literal token text and recovered entity spans for review.
- Enter input, ordered token start/end json, ordered labeled span json, tags json for reverse conversion.
- Choose the settings for token span label workbench, then run the tool.
- Review the result and any limitations before copying or downloading it.
Good to know
- 50,000 source characters, 2,000 ordered nonoverlapping tokens, 500 entities and exact original JSON integer offsets. Coordinates count Unicode scalar values, not UTF-16 units, bytes or grapheme clusters; combining marks remain separate.
- IOB2 and BILUO only, strict transitions and ASCII label names. Entity endpoints must be existing token edges. No NLP, automatic tokenization, nested/overlapping entities, label repair or spaCy runtime/importer claim.
- Processing is local. Output is limited to 1,000,000 text characters unless a stricter file limit is stated.
Common questions
Does an emoji use two offset positions?
A single supplementary-plane emoji uses one Unicode scalar position here. A joined or combining sequence can use several positions. Supply coordinates using that declared convention rather than JavaScript UTF-16 length.
Is this input sent to a server?
No. The application processes this input inside your browser tab and does not save it to persistent storage.
Keep going with related tools
All text →Unicode Code Point Converter
Convert text to U+ values or Unicode escape notation and decode valid scalar-value tokens back to text.
TextExplicit Text Range Redactor
Replace exact Unicode scalar ranges without splitting emoji or disclosing removed text in a change report.
TextExact Text Offset Coordinate Workbench
Migrate declared text ranges between UTF-8 byte, UTF-16 code-unit and Unicode scalar coordinates with exact quotes and boundary checks.
TextSpan Annotation Adjudication Workbench
Compare two flat span annotation sets on one exact text and export explicit human decisions with original range and label evidence.
EducationInterlinear Gloss Alignment Workbench
Review word and morpheme-separator alignment and export safe original interlinear teaching examples when the declared checks pass.