Register the original work
A creator registers an original text and stores a signed provenance record.
Text Provenance
Protect your written works. Prove they're yours.
ObraVera fingerprints your existing written works — manuscripts, articles, books — so you can prove provenance even after editing, reformatting, or conversion. No changes to the original needed.
The Problem
Forward-labelling doesn't protect existing works
Every day, published works are scraped into AI training sets. Metadata is stripped. Attribution is lost. Creators have no reliable way to prove a work is theirs once it leaves their hands.
Forward-labelling approaches embed metadata into AI-generated text at the point of creation. That helps identify machine output — but it does nothing for works that already exist. It cannot protect a manuscript written last year, or a novel published a decade ago.
The difference: forward-labelling marks what AI creates. ObraVera protects what humans created.
How ObraVera Is Different
We fingerprint what humans created. Not what AI generated.
Content Fingerprinting
Standards-track content identification
Cryptographic Signatures
Mathematical proof of authenticity
Resilient Matching
Identifies works even after significant editing
Multi-Format Support
TXT, DOCX, HTML, EPUB, PDF and more
Works retrospectively on existing published works — no changes to the original needed.
See It Work
The text is untouched. The provenance is still there.
ObraVera does not modify the original text. Instead, it computes a content fingerprint from the work as-is — using ISCC and related techniques — and registers it. Watch the scanner derive the fingerprint and match it against the record.
William stood on the branch, catching his breath. He felt a twinge in his stomach. He was hungry. It was past his mealtime. He wondered what his parents would think when they woke to find him gone. Would they go out searching for him? Perhaps they would think that he had fallen out of the nest during the night and been eaten. William felt sad and alone.
Status
READY
Content ID
—
Creator
—
Registered
—
Example Verification
See how ObraVera works without a live demo.
This is a real example from our internal test suite. The output below shows how ObraVera can link a reformatted or lightly edited file back to its likely source.
Compare a changed copy
We then verify a PDF version with UK/US spelling changes to see whether the content still traces back to the source.
Interpret the result
The result distinguishes exact identity from strong similarity, so an edited or reformatted file can still be linked to the original.
Test case
William Woodpecker, source text to edited PDF
Registered file: original `.txt` manuscript. Verified file: `.pdf` version with spelling changes. Result: still strongly linked to the same work.
- Registered source: clean original text
- Verified copy: PDF with UK/US spelling edits
- Use case: authenticity after formatting or editorial change
Real output
Verification summary
These values come from an actual ObraVera verification run. They show strong continuity despite file conversion and spelling changes.
Document identity
99.6%
Top source match
Likely source
ISCC status
No exact match, 62.8% similarity
What this means
The file is not an exact binary match, but the underlying work is still strongly identifiable. That is the point of ObraVera: verification that survives real-world editing and conversion.
Show technical detail
{
"match_outcome": "likely_source",
"document_identity_probability": 0.996,
"iscc_status": "none",
"iscc_similarity_score": 0.628,
"notes": "Edited PDF remains strongly attributable to the original text."
}Who It's For
Built for the people who create
Authors & Writers
Register your manuscripts, articles, and creative works before publication.
Publishers
Verify provenance across your catalog. Detect unauthorized reproductions.
Authors' Organizations
Protect your members' intellectual property with verified provenance.
Interested in text verification for your organization?
Join the early access list