Implemented · synthetic only
Request and scenario planner
A Python starter validates fictional research requests, requires one baseline and bounded intervention scenarios, and identifies work still needed. Predictions remain explicitly uncomputed.
Kira Labs · Health & anatomical research · early stage
Patient Horizon.
A record that evolves with evidence.
Patient Horizon is an experimental, clinician-oriented research project exploring how approved lab results, scans and observations could eventually update a versioned virtual patient record and help doctors review relevant published research. The long-term goal includes reviewed anatomical avatars and scientifically validated, carefully bounded scenario comparisons.
Current status — October 16, 2026: A synthetic-only Python request planner and research specifications exist. Live patient records, imaging intake, automatic literature search, functioning organs, treatment prediction and a clinical application are not implemented or validated.
Documented implementation · October 15–16, 2026
Built foundations.
Not a medical product.
Patient Horizon begins with a deliberately small, testable software foundation. Its present output is a plan for research, not a scan interpretation or a patient's predicted outcome.
Documented checks
4 starter tests passed
The October 15 local checks covered planning, duplicate-identifier rejection, baseline requirements and invalid time horizons. They do not test medical predictions or an avatar.
Engineering source, not integrated
Avatar Builder snapshot
A versioned, hash-recorded copy of generic Avatar Builder source is preserved for future integration. It is not a working patient-matched body or a validated anatomy pipeline.
Evidence boundary: No actual patient information has been imported. This project has no functioning disease-progress model, treatment-ranking engine, clinical dashboard, active medical search or live background monitoring.
Research direction · future workflow
From observations
to reviewable questions.
The project is designed to keep observed patient information, source-linked publications and hypothetical computer-model outputs clearly separate.
1 · Patient record
Update as results arrive
With authorization and review, a future system would track labs, imaging, dates, units, corrections and an ongoing medical timeline. Images would only change the avatar where reviewed evidence supports the change.
2 · Scientific evidence
Find relevant research
A proposed evidence service would retrieve dated publications and clinical-trial records, preserve citations and limitations, and let clinicians read or listen to summaries. No automatic search is operational today.
3 · Validated models
Compare defined scenarios
A future physician-reviewed simulation would use one independently validated disease model and endpoint at a time, showing assumptions, uncertainty and relevant evidence—not guaranteed treatment outcomes.
Clinical boundary: Patient Horizon cannot currently diagnose, predict progression or recommend treatment. Real-patient research would require authorization, domain specialists, validation, access controls and appropriate regulatory review.
Avatar Body · respiratory research
Breathing is a system,
not just an animation.
On October 16, KiraWorld and Patient Horizon documented the requirements for a possible lung–body–NewBrain connection. This is a defined engineering research target, not an achieved lung demonstration.
Future geometry test
Controlled breathing motion
Measure lung and chest movement, repeatability, units and restoration using an approved anatomy asset. A moving mesh alone would not prove airflow or gas exchange.
Future physiology
Airflow and circulation
Separately model diaphragm movement, airway mechanics, pressure, gas exchange and transport through the blood using explicit equations and reference data.
Future two-way interface
NewBrain and body signals
Test both directions with measured, timestamped messages, disconnected controls and a real receiver. A scripted command or transport acknowledgment alone is not brain–body integration.
Not yet built: No lung simulation, respiratory physiology, gas-exchange solver or two-way NewBrain–lung result has passed a test. A separate KiraWorld forearm experiment did verify repeatable movement and restoration of 512 sampled weighted skin vertices; that is a geometric body result, not a lung result.
University and clinical research collaboration
Separate experiments.
Shared expertise.
The first meaningful collaboration could test a synthetic patient-data workflow, anatomy validation or one tightly specified model endpoint—without making unsupported clinical claims.
References: Kira Labs Patient Horizon README and October 16 status / integration records; KiraWorld anatomy requirements. These are internal project documents, not independent clinical studies. The linked film is an illustrative research concept, not a product demonstration.