cm.about · independent · LLC

The clinical data stack,
collapsed into one platform.

Built by people who've spent careers running clinical trials — and who've felt every seam in the legacy stack.

§01 / 04 topic: thesis premise: collapse the stack 8 tools → 1 platform

Most clinical data work is
plumbing, not science.

A typical Phase III trial uses eight or more tools: an EDC, a coding system, a SDTM mapper, a SAE reconciler, a query manager, an analytics layer, a CDISC standards library, and a workflow engine. Each has its own user list, audit trail, integration team. The teams spend more time exchanging files than analysing them.

end-to-end vision →
stack/before · after collapse 8 tools → 1 platform user lists: 8 → 1
cm-thesis // stack_collapse integration teams: many → one audit trails: many → one premise

Legacy stack · 8 tools

EDCvendor lock
Coding (MedDRA / WHODrug)batch sync
SDTM mappermanual
SAE reconcilerspreadsheet
Query manageremail-driven
Analytics layerdelayed
CDISC librarypaid plug-in
Workflow engineJIRA-grade
collapses
⇒
into

ClinMaker · 1 platform

AI-assisted workflowshuman-in-loop
One canonical data modelend-to-end
One audit trailappend-only
One RBAC + 11 rolesMFA required
CDISC + USDM built inversion-pinned
15 vendor adaptersOID-keyed
CORE rules pre-exportblocks ERR
Inspection-ready7y retention
user lists collapsed: 8 → 1 audit trails collapsed: 8 → 1 integration teams: many → 1 → lock in weeks, not months
§02 / 04 topic: written_thesis why we built it

Why we built ClinMaker.

est. 2026 · LLC · independent

The cost of the legacy stack is not just budget. It's lock time, audit risk, and the inability to know — at any moment — how clean your data really is. By the time a sponsor realises a reconciliation issue went unresolved for three weeks, the answer is already a query backlog, a SDTM blocker, and a slipped submission.

ClinMaker exists to collapse that work into a coherent platform: one canonical data model, one audit trail, one role model, AI-assisted with human review gates at every decision. Every action is recorded. Every artefact is traceable. Every standard is enforced where data is collected — not patched in at submission.

We're building it module by module, in the open. Today: Protocol Authoring + Study Build are live. Ten more workflow modules are on the build path. The vision is end-to-end automation; we ship modules one at a time, each with the same audit posture and human-in-the-loop discipline.

We're an independent software company. Not part of a holding group, not VC-backed, not optimising for an exit. We build the platform, we use it ourselves, and we ship it under direct change control.

§03 / 04 topic: current_state 2 live · 10 roadmap early access

Where we are today.
Two modules. Real customers.

We don't claim a finished platform. We claim two live modules, a clear roadmap for the rest, and the same audit + GxP posture across everything we ship — current and future.

see modules →
§01 live
Protocol authoring

USDM-native, M11-aligned, DDF-conformant. AI co-pilot drafts every section; humans approve. CORE rules engine blocks ERROR-severity exports.

standards · DDF · USDM 4.0 · ICH M11 · NCI CT
§02 live
Study build

USDM definition (or PDF, or USDM JSON) → CDASH casebook. Forms, fields, codelists, edit checks. Deploy to Rave or Veeva.

intake · USDM · PDF · USDM JSON · output · CDASH 2.2

The other ten workflow modules — ingestion, validation, queries, coding, reconciliation, SDTM/ADaM, analytics, patient profile, audit trail, user admin & help — are under development. We don't commit dates publicly. We commit milestones. Early-access pilots may include scoped previews of in-development modules.

§04 / 04 topic: principles 6 hard constraints enforced platform-wide

What we won't ship.

A short list of constraints we hold ourselves to. These aren't aspirational — they're hard rules on what reaches our customers.

ask us about these →
principles/index 6 constraints enforced platform-wide → verified per release
cm-principles // inventory scope: every module exceptions: none tested
P-01 No unaudited mutation. If a value changed, the audit trail explains who, what, when, and why. There are no exceptions, no exempt modules, no batch-write bypasses.
P-02 No black-box AI. Every AI decision surfaces inputs, confidence, and rationale. Humans approve or reject; the system records the decision. AI rows are distinguishable from human rows in the audit.
P-03 No standards as afterthought. CDISC, USDM, MedDRA, WHO Drug are wired into the data model — not bolted on at export. Standards version pin per study.
P-04 No deferred compliance. GxP isn't an enterprise tier you upgrade to. 21 CFR Part 11 posture is the floor for every customer, every study, every environment.
P-05 No vendor lock-in. Export your data, your audit trail, your mapping specs at any time. Open formats, open standards. Your data is yours, always.
P-06 No fake "AI-powered". If a workflow runs deterministically, that's what we ship. AI is for problems that genuinely need judgement — not for decoration.
cm.demo

Curious to talk?
30 minutes. Live platform.

We'll show you the working platform — not a slideshow. Working session, not a deck.

Talk to us → or tour the platform