Unlock archives together
Transkribus combines AI-powered text recognition with collaborative workflows. Engage volunteers, students, and researchers to transcribe, correct, and enrich historical documents at scale.
Why AI-assisted crowdsourcing is faster
Traditional crowdsourcing asks volunteers to transcribe from scratch. With Transkribus, AI does the heavy lifting first — volunteers correct rather than create.
Traditional Crowdsourcing
Recruit volunteers
2–3 monthsManual transcription from scratch
12–24 monthsReview & correct
3–6 monthsFinal quality check
1–2 monthsWith Transkribus
AI pre-transcribes all pages
HoursVolunteers verify & correct
3–6 monthsBuilt-in quality control
OngoingThe platform behind every crowdsourcing project
Volunteers transcribe directly on the document image. Layout analysis, text recognition, and quality tools work together to make contributions accurate and efficient.
Platform comparison
Transkribus vs. other crowdsourcing platforms
How Transkribus compares to FromThePage, Zooniverse, and similar crowdsourcing tools for handwritten document transcription.
| Feature | Transkribus | FromThePage / Zooniverse |
|---|---|---|
| AI pre-transcription | Built-in HTR produces a first draft at 70–95% accuracy — volunteers correct rather than transcribe from scratch | No AI assistance — volunteers transcribe every word manually |
| HTR model training | Train custom AI models on your specific handwriting with as little as 5,000 words of training data | Not available — no machine learning capabilities |
| Role-based access | Owner, editor, and viewer roles with granular permissions per collection | Basic user roles — limited control over who can edit what |
| Progress tracking | Dashboard shows completion status per page, per document, and per collection | Varies — some platforms offer basic progress indicators |
| Full-text search output | Every transcribed page is full-text searchable. Publish as a searchable digital edition via Transkribus Sites | Search depends on export format and external tools |
| EU-hosted | All data processed and stored in Austria (EU). GDPR-compliant. No third-party cloud dependencies | Typically US-hosted — data sovereignty varies by provider |
Comparison reflects publicly available feature sets. Platforms may offer additional capabilities not listed here.
Built for collaborative transcription projects
Everything you need to manage volunteers, track progress, and maintain quality across large-scale transcription efforts.
Role-based access
Assign owner, editor, or viewer permissions to each team member. Control who can edit, review, or just browse your collection.
Progress dashboard
Track completion across your entire project. See which pages are done, in progress, or still waiting for review.
AI pre-transcription
300+ public AI models produce a first draft in seconds. Volunteers start with text already on the page — not a blank screen.
Custom tag sets
Define your own tags for people, places, dates, or any entity type. Volunteers enrich the text with structured annotations.
Quality metrics
Confidence scores and progress indicators are available and under active development. Reviewers can use these signals to focus on areas where the AI is least certain.
Export formats
Export to PAGE XML, ALTO XML, TEI, plain text, DOCX, or searchable PDF. Integrate results into any archival system.
Built for collaboration
Manage your project with Transkribus Projects
Success story
25,000 Wills Transcribed: How AI Accelerated a Major Crowdsourcing Project
Launch your crowdsourcing project with AI support
Speak with our team about setting up a collaborative transcription project, or create a free account to evaluate Transkribus with your own materials.
Organisation plans available for institutional projects