10s of K
Instruments per hospital, every day
Each one must be cleaned, inspected, packed, sterilized, traced, and matched to the right set, every cycle. The decontamination loop never stops.
// identification system
TRL 3 · early validation
A fingerprint for
surgical instruments.
We identify reusable instruments by their intrinsic physical signature. No laser engraving, no chips, no tags. Place it down, we recognize it.
- ≥99.95%accuracy goal · 98.9% at MVP today
- ~3sper instrument, assisted “place & go”
- 0marking, chips, or tags required
01
The problem
Sterile processing runs on manual reconciliation.
A hospital CSSD cycles tens of thousands of instruments, every day. The cycle runs operating room → collection → wash → identification → sterilization at 134 °C. Today, that identification step is still done by hand, one instrument at a time. That is the bottleneck.
15–30s
Per instrument, identified by hand
Manual identification is step four of the cycle, and its slowest. Each item is checked and logged individually, which is exactly where throughput stalls.
missing
or extra
or extra
Set discrepancies cost the most
Interviews with CSSD leadership confirm the costly failure mode isn't unreadable barcodes. It's incomplete sets and weak accountability at handoff.
Why the existing options fall short:
DataMatrix / barcodesRequire direct marking, line-of-sight, and handling effort per item.
RFIDNo line-of-sight, but adds per-instrument tag cost, attachment burden, and durability concerns.
Camera AIRecognizes types, not individual units; can't distinguish wear, cycles, or repair history.
ManualAlready the baseline. Labor-intensive, inconsistent, weak accountability.
02
How it works
Place the instrument. Get an identity.
Our hardware captures the instrument's intrinsic physical signature. A trained model maps that signature back to a device record in about three seconds, no marking required.
-
01
Place
The instrument is placed in the station during normal inspection or packing, with no extra handling step.
-
02
Sense
The hardware captures the instrument's intrinsic signal signature.
-
03
Classify
A neural classifier matches the signature to a record in the device library. Instrument type is validated today; individual-unit identification is in early validation.
-
04
Verify
Workflow software confirms set completeness, flags discrepancies, and logs the result into the CSSD's existing documentation system.
Product architecture
Four layers, one workflow.
L1
Sensing hardware
Passive, contactless capture hardware, engineered for repeatability under wear and sterilization cycles.
L2
ML classification
A trained classifier converts raw signal into instrument identity, built on proprietary, sterilization-conditioned datasets.
L3
Workflow software
Completeness checks, discrepancy alerts, exception handling and logging, designed for a ~3-second operator interaction.
L4
Integration layer
Minimum-viable interfaces to existing CSSD documentation, tracking platforms (T-DOC, Steelco, Belimed), and structured exports.
03
Why now
Regulation is forcing the question. Marking can't always answer it.
EU MDR requires a Unique Device Identifier on every reusable instrument, surviving every sterilization cycle. The EUDAMED database becomes mandatory in 2027. For instruments where direct marking is infeasible or operationally too slow, a validated marking-free alternative moves from nice-to-have to necessary.
May 2017
EU MDR enters into force
Regulation (EU) 2017/745. UDI required on the device itself for reusable medical devices, with exemptions only where direct marking interferes with safety or performance.
Ongoing
UK MDR & CQC traceability
CQC inspections explicitly expect “full traceability from decontamination to point of use.” UK CE/UKCA transition runs through 2030.
2027
EUDAMED becomes mandatory
Use of the EUDAMED database becomes mandatory, with UDI/Device registration obligations applying to devices on the EU market. That turns UDI compliance from paperwork into an enforced, queryable record.
The gap
“Legacy” instruments can't be marked
Hospitals hold thousands of older reusable instruments that cannot be retro-marked without enormous cost or compromising the device. This is exactly where a marking-free identity is irreplaceable.
04
The market
Underpenetrated. Growing. Regulated.
We don't need to displace a mature market. Most European CSSDs still rely on partial or manual tracking. The opportunity is to fit the gap, not replace the entire stack.
€85M
SAM · European surgical-instrument tracking market, annual
€290M
TAM · global market for the same category
13%
CAGR · growth driven by EU MDR regulation
~8K
EU hospitals without a functional tracking system
Beachhead segments
Beachhead
Large EU hospitals under MDR pressure
High throughput and thousands of unmarked “legacy” instruments that can't be retro-marked without enormous cost. This is where a marking-free identity is irreplaceable, and our first target.
Primary
Modern / automated CSSDs
Already redesigning workflow around robotics and integration, and the most open to a marking-free station.
Primary
Outsourced CSSD providers
Single contract can unlock multi-site deployment and produce reference data fast.
Secondary
Instrument OEMs
Aesculap, Stryker, J&J: compliance and differentiation logic; OEM licensing later.
Channel
CSSD equipment vendors
Belimed, Steelco, Getinge / T-DOC, STERIS: integration partners or distribution channels.
Adjacent
Incoming-inspection & counterfeit detection
Aerospace and medtech manufacturing: same sensing concept, high willingness to pay; reserved as a medium-term expansion.
05
Versus alternatives
We don't replace tracking. We close a gap inside it.
The strategic position is precise: complement existing tracking platforms by solving completeness verification and identification for items that aren't practically markable, or where line-of-sight scanning kills the workflow.
Approach
No marking required
Individual identity
Workflow fit
Per-item cost
Object Fingerprint
Yes
Type now
~3s
€0
DataMatrix / barcode
No
Yes
Line-of-sight
Low
RFID tag
No
Yes
Bulk read
Tag cost
Camera AI
Yes
Type only
Visual
Zero
Manual visual check
Yes
Limited
Slow
Labor
We're honest about the trade-off: for bulk reading of many tagged items at once, RFID is still faster. Our wedge is the items that can't be marked, or where per-item, no-contact identity matters more than bulk throughput.
06
Roadmap & traction
24 months. Three milestones.
We treat the company as a validation-stage business. Current readiness is TRL 3: a working MVP proven in the lab. From here the plan is 24 months to CE marking and the first hospitals, in three clear milestones.
M1 · 3–6 mo
Lab validation of stability
- Hundreds of autoclave cycles
- Temperature, chemistry & wear stress
- Signature repeatability proven
M2 · 6–12 mo
Proof-of-concept pilot
- Narrow pilot at a partner hospital
- A set of “legacy” instruments
- Integration with existing CSSD software
M3 · 12–24 mo
Full launch · CE · 5 hospitals
- CE certification
- Sale to five hospitals
- First OEM partnership
Targets & readiness
≥99.95%
Target identification accuracy
98.9%
MVP accuracy today, on instrument type
~3s
Read time per instrument (assisted)
TRL 3
Current technology readiness
24 mo
To CE + first hospitals
0
Marking, chips or tags required
Interest
We're in early conversations with several hospitals and a medical-device manufacturer who see the need and want to follow our progress. It's genuine, encouraging interest, and we're deliberately keeping it at that: nothing here is a signed deployment yet.
07
Team
Cross-domain by design.
The defensible asset is the combination of sensing hardware, signal processing, applied ML, and CSSD workflow knowledge, held by the same small team.
Jaroslav Marek
CEO
Medtech leader across sales, procurement and research. CTU FBMI roots in signal-based recognition of surgical instruments, the scientific backbone of the platform. Owns commercial strategy and clinical relationships.
Dominik Fischer
CFO
IT, business and finance background. Owns the financial model, fundraising and commercial operations, and helps steer go-to-market and partnerships.
Radek Daniel Paseka
CTO
Hardware expert with deep-learning research chops. Owns sensing hardware, the ML fingerprint model, and the path to certification.
Talk to us
Looking for one design-partner CSSD.
We're scoping the first on-site pilot. If you lead a CSSD or sterile-services team, or if you fund deep-tech medical devices, we'd value 20 minutes of your time.