Language research continues privately
The parser is an internal research tool. Visitors cannot submit sentences through this website.
Kinesema approach · Language before motion
Sign languages are not visual versions of spoken languages.
Sign languages have their own grammar, use of space, rhythm, expression, and cultural context. Kinesema therefore begins with language structure rather than replacing English words with signs or generating motion immediately.
Why structure comes first
Structure before expression.
An avatar cannot repair weak or incorrect language structure. Kinesema separates the stages so each transformation can be inspected before it becomes motion.
- Input English input
- Private development Language analysis
- Research prototype Proposed SgSL structure
- Required Human review
- Later stage Motion-ready representation
- Later stage Visual motion
Language analysis and proposed representations are being developed privately. The public open-hand study is a separate form experiment; it does not turn parser output into signing motion.
Inspectable transformation
Show the transformation, not only the result.
A confident interface should not make an uncertain output look authoritative. Kinesema’s white-box direction keeps the path available for inspection.
- what the current system recognised;
- what it proposed rather than presenting a final translation;
- which parts remain uncertain;
- where human review is still required.
The research direction and public boundaries are explained in the Kinesema documentation.
Human and community judgment
Human communication stays central.
Kinesema is intended to support access, learning, communication, and future motion review. It is not a replacement for Deaf people, interpreters, educators, signers, or community authority. Language- and culture-sensitive decisions require relevant human review.
Read the separate accessibility commitments and current limitations.
Current boundary
What Kinesema currently does.
The output remains inspectable
Proposed representations and their uncertainties are inspected inside the lab before any claim of accepted translation.
Human review is required
The current result is a reviewable proposal, not linguistic, cultural, or community authority.
Motion is still being developed
Reviewed motion plans, avatar signing, and a final translation remain outside the current public stage.
Still learning
The approach will continue to learn.
Kinesema remains early-stage work. Feedback from Deaf people, signers, educators, linguists, accessibility practitioners, and technical reviewers helps reveal where the system must improve.