Patient-specific implants take 5-10 days.
Bespoke services route through engineering review per case. Time-to-implant is a planning constraint, not a clinical one — and it pushes surgeons toward shelf solutions when the case can't wait.
For investors · pre-seed conversations
Cranioplasty planning, made reviewable.
A research-stage AI-assisted workflow for personalized cranial reconstruction. Four modules a clinician can inspect — defect, boundary, implant, engineering — not a black box. The MVP is shipping; we're in active pre-seed conversations to land the next phase's data and regulatory scaffolding.
- Reviewable demo
- Live
- Six fixed-tail SkullFix research cases on the demo
- Model development
- Closed arc
- Honest evaluation; binding constraint named
- IP path
- In flight
- Monash Innovation IDF channel
- Round
- Pre-seed
- Active conversations
Problem
Cranioplasty planning hasn't caught up to the scan.
CT and 3D printing changed what's possible in the operating room. Planning workflows didn't follow. Three friction points keep cranioplasty bespoke, slow, and hard to review.
DIY planning is ad-hoc and undocumented.
When the bespoke route is too slow, surgeons fall back to manual templates, intra-op shaping, and informal notes. No structured engineering feedback, no audit trail, no reproducibility for the next case.
AI tools that don't explain themselves don't ship.
Autonomous-output models struggle with clinician trust and regulatory pathways. Cranioplasty needs review, not opacity — and most current AI tools are built the wrong way around for that.
Wedge
AI-assisted, reviewable, not autonomous.
CranioSwift is designed around the review step, not against it. Four modules with clean interfaces. Every output a human can inspect. Built for the workflow surgeons already use, not a workflow we wish they used.
Four modules, four audit points.
A defect recognition, B boundary constraint, C implant shape, D engineering feedback. Each module produces a reviewable artifact a clinician or engineer can sign off on independently — not a single end-to-end black box.
Fast turnaround the workflow expects.
Target turnaround: 24-72h from CT to reviewable implant candidate, not 5-10 days. The slow step today isn't computation — it's coordination. We give clinicians and engineers a shared surface to compress it.
What we ship vs. what we don't.
Today's shipping bundle is a reviewable research-line demo on six fixed-tail cases. Promote-grade commercial performance needs private clinical data we don't yet have. The roadmap below names the gap explicitly; the pre-seed round closes it.
Proof · invite-only research demo
Six tail cases. Click through and judge.
The /demo/ workbench shows the shipping demo bundle's actual output on six fixed-tail public-cranial-dataset cases. You can rotate the implant candidate, inspect the four-module evidence cards, and see the per-case quality numbers. What you see is what we have today — no demo magic, no hidden cases, no inflated metrics.
Enter invitation code →
No code yet? Request an investor access code
Traction
What's already in place.
Four anchors today, all independently verifiable. We surface them so investor diligence can confirm the foundation before any conversation about the next phase.
Shipping product
Reviewable demo, live
Six fixed-tail SkullFix research cases on the public demo, with watertight surfaces and zero skull-overlap voxels. Inspect the actual output at cranioswift.com/demo/.
Research-line bundle, demo-grade; commercial-grade bundle on the Phase 8 roadmap.
IP path
Monash Innovation IDF in flight
Confidential method-level disclosure routed via the Monash Innovation IDF channel to institutional patent counsel. Public-facing language stays at outcome / module-name level until the institutional provisional is filed.
Filing route: institutional (Monash Innovation IDF)
Honest evaluation
Closed model-development arc, binding constraint named
A series of bounded research experiments closed with a precise diagnosis: the next step toward commercial-grade performance requires private clinical data scale, not further architecture work.
Public summary: /phase-summary/ · full per-phase detail under NDA.
Surface posture
Research-use software, disclosed up front
Every page carries a persistent research-use banner. Outputs are clearly framed as not-for-clinical-decision-making during the research phase. This posture is the right one for the regulatory path we'll walk.
See banner at the top of every page
Roadmap
Past, now, next.
One timeline across model development, product, IP, and clinical / regulatory. No phase started before its predecessor closed; each phase named what it learned, including the things that didn't work.
-
Past · research arc closed
Bounded research, honest evaluation.
- Shipping bundle Reviewable workflow baseline live on the demo
- Bounded experiments Iterative research closed with a precise diagnosis on the constraint for commercial-grade performance
- Binding constraint Private clinical data scale, not architecture or loss design
- Close ML architecture work paused pending data + funding
-
Now · product surface + parallel tracks
Surface + fundraising + parallel data conversations.
- Product surface Investor view, methodology page, clinician demo enhancements
- Parallel Private clinical data acquisition conversation underway (director-led)
- Pre-seed Active angel conversations
- IDF Monash Innovation IDF in flight; institutional provisional patent pending validation gate
-
Next · post-funding
Private clinical data + engineering feedback + regulatory engagement.
- Next ML phase Resume model development on private clinical data, targeting commercial-grade performance on held-out clinical-distribution cases
- Module D Engineering-feedback upgrade toward patient-specific structural analysis
- Regulatory Regulator early conversation post-provisional; clinical pilot site under appropriate institutional frameworks
- Surface Owner-only live workbench upgrade; additional partner-facing surface scoped as needed
Use of funds
Four buckets, one binding constraint.
The pre-seed round funds four work streams in parallel. The bucket split, round size, and milestone-tied schedule are in the investor materials packet (under NDA / first conversation).
Private clinical data acquisition
Institutional framework with at least one clinical partner site — anonymization, secure storage, and the data engineering to ingest paired cranial CT + defect-implant cases at clinical distribution. Multi-week calendar; the binding constraint for the next ML phase.
ML R&D on private data
Next-phase model development on the private corpus, aimed at lifting held-out clinical-distribution performance past the commercial-grade promote bar. Engineering-feedback module upgrade in parallel.
Regulatory + IP + legal
Monash Innovation IDF follow-through into institutional provisional filing. AU regulator early conversation scoping. Legal review on data licensing posture and clinical-pilot framework.
Operational buffer
Reserved for unplanned data-acquisition friction, longer-than-expected institutional review, or a tactical hire if the round closes at the upper end.
Team
Who's building this.
Lean founding setup, Monash-anchored, with named open slots for clinician and operator advisors we're actively talking to. We name the gaps because investor diligence will find them anyway, and they're addressable.
Michael Peng
Founder · lead engineer
Monash researcher building the four-module pipeline end-to-end. Runs the model-development experiments, ships the product surface, owns the IDF channel and the fundraising track.
Monash University · Addin (planned spin-off vehicle) · MCAM
Scientific advisor
Cranial reconstruction research line
Senior Monash academic anchoring the scientific direction, the IDF channel, and the publication path. Named in confidential investor materials; placeholder on this public surface until confirmed.
Monash · cranial / craniofacial research line
Clinician advisor (open)
Cranioplasty workflow + clinical pilot
AU-based neurosurgeon or craniofacial surgeon with cranioplasty volume and an interest in workflow tooling. Active conversations; introductions welcome.
Open slot
Operator advisor (open)
AU med-tech regulatory + commercialization
AU med-tech operator with regulatory pathway experience, ideally with prior surgical-tool or implant company involvement. Active conversations; introductions welcome.
Open slot
Regulatory path
Monash IDF → provisional → TGA early → FDA TBD.
The IP and regulatory milestones run in sequence, not in parallel. Each pin below has a clear next-action and a clear gating dependency.
-
Active
Monash Innovation IDF
First-round disclosure package routed through the Monash Innovation channel.
Next: institutional review and iteration.
-
Scheduled
Institutional provisional patent
Method-level provisional filing routed through Monash Innovation to patent counsel; the subject method is described under NDA in the investor materials packet.
Next: complete the validation gate; provisional draft follows.
-
Planned (post-funding)
AU regulator early conversation
Scoping conversation on the appropriate classification and clinical evidence requirements for an AI-assisted reviewable cranioplasty workflow.
Next: scoping memo once private clinical data acquisition is underway; conversation initiated post-funding.
-
Future (out of round)
International regulatory pathway
Out-of-scope for the pre-seed round; re-scope at the next round when AU clinical-pilot evidence is available.
Next: defer.
How we compare
Three reference points.
vs Bespoke PSI services
Speed + reproducibility, not custom-engineering scale.
Materialise PSI and similar fully-bespoke services route every case through case-specific engineering. 5-10 day lead times. We target 24-72h with the structured engineering review built into the workflow itself, not bolted on per case.
vs DIY craniotomy + manual planning
Audit trail + structured handoff, not ad-hoc.
Today's fallback when bespoke is too slow: surgeon judgement + intra-op shaping + informal notes. Hard to review, hard to reproduce, hard to teach. We give clinicians and engineers a shared workspace with provenance per case.
vs autonomous AI tools
Reviewable by design, not opaque-then-explain.
Most current AI implant tools optimize for end-to-end output and add explainability post-hoc. We optimize for the review step from the start. Four reviewable modules, each with a sign-off interface, each independently improvable.
Next conversation
Let's talk.
Pre-seed round, active conversations. We're open to introductions, diligence questions, and friendly-investor demos. The investor-materials packet (full deck, data-acquisition plan, regulatory scoping memo, use-of-funds detail) goes out on request.