Built on a governed foundation
OpenTrials inherits the OpenLI Codex governance posture: 7-tier RBAC, audit trails on every agent action, human approval gates, PII sanitisation hooks. Aligned with ICH E6(R2/R3) GCP and 21 CFR Part 11 from day one.
LIVE DEMO · v0.3.17 · CT.OPENLI.AI
AI-native RBQM and central statistical monitoring for the modern clinical trial.
OpenTrials is a live OpenLI product. Try the demo at ct.openli.ai — 8 role-based logins, 5 demo studies preloaded (PHOENIX-3 plus 4 Vanguard packages), full CDISC SDTM 3.3 / ADaM 1.3 import paths, and a SHA-512 audit trail underneath every action.
OpenTrials brings the OpenLI agentic AI foundation to clinical trial Risk-Based Quality Management. Built on the same OpenLI Codex runtime that powers our healthcare integration products — with the same governance, audit trail and dual-runner architecture, specialised for the clinical trial sponsor and CRO buyer.
Live at ct.openli.ai
21 CFR Part 11 · ICH E6(R2/R3) · GAMP 5
CDISC SDTM 3.3 / ADaM 1.3
Built on OpenLI Codex
Why OpenTrials
Clinical trial Risk-Based Quality Management is shifting from spreadsheet-driven site monitoring to real-time risk surfaces. AI-assisted RBQM is the obvious next step, but the regulated nature of clinical trials means agentic systems must come with governance built in — not bolted on. That is exactly the OpenLI Codex thesis. OpenTrials applies it to the pharma sponsor and CRO buyer.
The same security review
Pharma buyers who already use OpenLI HIE, IMX Monitor or Integrai have already done the OpenLI Codex security review. OpenTrials inherits that posture — one less procurement gate to clear.
Live today, evolving
v0.3.17 ships RBQM lifecycle, study-package import, statistical engine, signal review and the SHA-512 audit trail. v0.4.x adds the EDC connector wave (Medidata Rave Architect XML adapter as proof-of-generality), then submission packaging.
What ships today (v0.3.17)
The capabilities below are live in the demo environment. The GitHub repository carries the canonical release notes.
RBQM lifecycle
KRI definition, central monitoring, signal review and disposition with full role separation between data managers, medical monitors and biostatisticians.
3-tab study-package import
Folder picker, ZIP upload (in-browser jszip extraction), or server-folder selector. CDISC SDTM 3.3 / ADaM 1.3 datasets land with full audit lineage.
Statistical engine
Bayesian KRIs, central tendency, funnel plots, enrollment anomaly detection. 32 pre-built tests in the Test Registry, all backed by closed-form pytest evidence.
Vanguard methodology coverage
OpenTrials engine + signal_type + kri_category vocabulary covers the 12 cluster families from the Vanguard cluster_summary_combined_final_158 reference set.
Multi-tenant + per-study RBAC
Sponsor / CRO / Regulatory / Site organisation types. Per-study access matrix at /admin/study-access. Time-boxed inspector access for regulators.
SHA-512 audit trail
Append-only hash chain on every mutation. 21 CFR Part 11 Subpart B + EU Annex 11 alignment with 15-year retention. Inspector view available to read-only viewers.
Cost gate + dual-runner
Mock-mode runners by default. Toggle real OpenAI / Claude tokens at /settings/runners via a sentinel-file (no service restart). TopNav cost meter shows live spend.
5 demo studies preloaded
PHOENIX-3 (Phase III NSCLC, 387 subjects, 8 KRIs, 6 signals) plus 4 Vanguard packages (ABC123 + Enterprise v4.4 + CTMAS v4.4 + Submission v4.4-define).
On the v0.4.x roadmap
EDC connector wave (Medidata Rave Architect XML), JSONB sidecar typed-hot-path performance work, Define-XML 2.1 submission packaging UI. See the GitHub release train.
FOUNDATION
Built on OpenLI Codex
OpenTrials is built on the same OpenLI Codex foundation that powers every shipping product on the OpenLI family tree. The runtime, the governance framework, the audit trail, the dual-runner architecture — all already in production today across our healthcare integration cluster.
LIVE DEMO
Try OpenTrials right now.
Open ct.openli.ai. 8 demo logins are documented (one for each role — superadmin, study admin, data manager, medical monitor, biostatistician, CRA and viewer) with 5 preloaded studies. Source on GitHub.