Terminus M4
Quality decisioning
A temperature excursion assessed against the product's own stability model, in under a minute, with the trail an inspector will ask for.
See the application →Applications for regulated pharma and logistics
Software for the decisions your inspector will ask about.
Synlogica builds the applications regulated teams use every week — quality decisions, controlled documents, environmental monitoring, freight invoices, training, laboratory planning, transport and supplier negotiation. Each one runs on your own tenant in the EU, with versioned records, electronic signature and an audit trail an inspector can follow.
19 yrs
in pharma quality, logistics and procurement
EU
your own tenant, hosted in the EU
0 LLMs
in any decision loop
BUILT IN EU · GDPR · 21 CFR PART 11 · EU GMP ANNEX 11 · EU GDP · ISO 27001 & 42001 ROADMAP
Every application says which stage it is at ↓
The applications
The applications
Where the same decision comes back every week
At that point it stops being a consulting question and becomes a system. Every application here carries the same controls — versioned records, e-signature, an audit trail an inspector can follow — and runs on your own tenant in the EU. They are not all at the same stage, and each one says which.
SQH
Documents & quality system
Controlled documents, training, deviations, CAPA and change control on one process engine, with e-signature and a hash-chained audit trail.
See the application →SEM
Environmental monitoring
Water quality against regulatory limits, with laboratory reports read straight out of the PDF instead of retyped.
See the application →REFRAKT
Freight invoice pre-audit
Every freight invoice line checked against the carrier contract before the payment run, not recovered after it.
See the application →Skillon
Training & competency
Who was trained on which version of which procedure, what expires when, and the proof it was complete on the day an inspector asks.
See the application →
QCPlanner
QC laboratory planning
Sample intake from LIMS, then a constraint solver that schedules the laboratory against real capacity, analyst competency and predicted OOS risk.
See the application →
Terminus M1
Transport decisioning
Freight tenders scored on landed cost rather than headline rate, with the leakage between award and invoice visible while the round is still open.
See the application →
Terminus M5
Negotiation decisioning
Supplier rounds run as multiple equivalent simultaneous offers, with your real BATNA stated in time and inventory rather than assumed.
See the application →
GRIP Risk Intelligence
Transit & geopolitical risk
Corridor and transit exposure for cargo and marine underwriting, read against a book of business rather than against headlines.
See the application → J-lab Illustrative
J-lab
Laboratory execution (ELN/LES)
A laboratory notebook and execution platform in development — methods, sealed execution, typed results, independent review.
GRIP · the engine under the hood
Hard rules, probabilistic models, one decision policy
Every recommendation carries the exact rule and model versions that produced it — reproducible bit-for-bit, and packaged so a reviewer or inspector can reconstruct what was decided, on what evidence, under which ruleset.
01 · Hard rules
Deterministic, versioned thresholds and policies — approved limits, storage classes, quality-agreement clauses. Tenant-configurable, never silently changed.
02 · Probabilistic models
Arrhenius kinetics, Bayesian elasticity, Monte-Carlo coverage — each with a stated confidence interval and the data-quality flags behind it.
03 · Decision policy
Maps the analysis to an advisory recommendation — and hands it, with its evidence, to the authorized human who makes the binding call.
New to these terms? Read them in plain language
- Arrhenius
- A formula from chemistry: the warmer a product gets, the faster it spoils. Terminus M4 uses it to work out how much shelf life an excursion really cost.
- Monte-Carlo
- Instead of one guess, the computer plays out thousands of possible scenarios and reports how often the result stays safe.
- MKT
- One single “average” temperature that fairly sums up a whole journey of ups and downs.
- MESO
- Putting several equally good offers on the table at once — the supplier's choice reveals what they really care about.
- BATNA
- Your best plan B if the deal falls through. Knowing it tells you when to say no.
- ALCOA+
- The regulator's checklist for records you can trust: who wrote it, when, is it the original, is it correct.
Data protection
GDPR is built into the applications, not only into this website
A regulated customer is asked the same questions about every supplier: where does the data live, who can reach it, what a language model sees, and how a record is deleted. These are the answers, and each one names the application it comes from.
All applications
Your own tenant, in the EU
Every application runs on a tenant dedicated to your organisation, hosted in EU regions (AWS Frankfurt, Hetzner Falkenstein). Deployment on your own infrastructure is available where policy requires it.
REFRAKT · QCPlanner
Tenant isolation enforced in the database
Row-level security in PostgreSQL is the backstop behind every query, and the tenant is stamped on every audit row — isolation does not depend on the application code remembering to filter.
REFRAKT · Terminus M4 · SEM · QCPlanner
Language models read, code decides
A model only extracts documents — a stability PDF, a carrier contract, a laboratory report. Every number is computed in code, every decision is signed by a person, and no model is ever trained on your data. QCPlanner runs its model locally, so batch data never leaves the server.
REFRAKT
Erasure that reaches storage
Retention is configured per organisation. Erasure removes the data from the database and object storage, and per-organisation encryption keys let raw model input and output be crypto-shredded. The erasure path is covered by an automated test.
SQH
DSAR and ROPA inside the quality system
Data-subject requests, the record of processing activities, legal requirements and interested parties are records in the compliance module — with the same workflow, e-signature and audit trail as a deviation.
SQH · Skillon · Terminus
Append-only audit trail, password-bound signature
Audit entries are hash-chained and cannot be edited; Skillon enforces immutability with a database trigger, SQH exposes a verification endpoint. Every controlled step needs an electronic signature with password re-authentication.
And this website
Cookieless analytics, no third-party marketing trackers, no embedded widgets that profile you. The booking link opens Calendly in a new tab and sets nothing on our domain. Data-subject requests are answered within one calendar month.
Compliance & trust
Built for the regulator, not just the demo
Every decision is built to answer the questions an inspector will really ask. Hosted in the EU. Backed by evidence. Mapped to the standards you already work with.
- Built in EU
- EU data residency
- GDPR
- GAMP 5 Cat 4
- 21 CFR Part 11
- EU GMP Annex 11
- EU GDP
- Cookieless analytics
Integrated Management System (IMS) — certification roadmap
See the full trust pack → - ISO 9001 · Quality
- ISO 27001 · Information security
- ISO 42001 · AI management
Talk to us
Start with your own decision package
Working session
Validate against your SOPs, on your data
We start with a working review of a sample decision package and its validation assumptions against your intended use — to agree the configuration baseline and the shared-responsibility split before anything goes live.
GAMP 5 Cat 4 documentation set & supplier evidence
Configurable thresholds, mailing groups & hosting
EU data residency or alternative cloud / on-prem
Bring a real workflow — a recent excursion, an upcoming tender, or a supplier price claim — and we walk it through together.
Book a call →
Email Adam directly — adam.karpinski@synlogica.ai
FAQ
The questions QA & validation always ask
Is this a batch-release system?
No — it is advisory by design. Terminus M4 produces a decision-support recommendation (RELEASE / CONDITIONAL / HOLD) with full traceability; the formal batch disposition is made and signed by the Authorized Quality Reviewer (QA / QP), recorded in your QMS.
How is it validated?
Terminus M4 is positioned as configurable software (GAMP 5 Category 4), validated through a risk-based lifecycle — intended use, configuration, requirements traceability, IQ/OQ/PQ, data-integrity and audit-trail verification. Synlogica provides product and supplier evidence; the regulated user owns the validated state for their intended use.
What about data integrity and 21 CFR Part 11?
Records are versioned; controlled edits require a reason and an e-signature gate (password re-authentication); the audit trail is immutable, attributable and exportable as CSV. Sensor data is screened for outliers, out-of-range readings and gaps before use (ALCOA+).
Which decisions does it cover?
Three domains on one engine: Quality (temperature excursions and shelf-life impact), Transport (tender leakage and freight risk), and Negotiation (MESO framing for API procurement).
Where does it run?
Cloud-native — Kubernetes on AWS EC2 with AWS S3 for durable, WORM-capable storage, under the AWS shared-responsibility model. EU data residency or an alternative cloud / on-prem topology is configurable per your hosting policy.
Does it use AI / LLMs to make the decision?
No. LLMs assist only with extraction and structuring of unstructured inputs (e.g. parsing a stability PDF or an unknown logger layout), always confirmed before use. The recommendation itself comes from deterministic, version-pinned rules and the kinetic / probabilistic models.
Load the evidence. Let the engine conclude. Keep the human in charge.
A full trail from logger file to client report — for quality, transport and negotiation.