Public overview — source code is private. This repo describes what the product does and how it's built.
Pharma regulatory and pharmacovigilance (PV) teams run document-heavy, deadline-critical, highly regulated workflows — RMPs, PSURs/PBRERs, PSMFs, signal management. The work demands deep expertise and airtight traceability. Generic AI tools can't be used because they aren't validated and leave no audit trail.
Praxara brings AI agents into regulated PV/regulatory work — built for GxP from the ground up:
- Agentic workflows for core regulatory deliverables (RMP / PSUR / PSMF and related documents), grounded in current regulatory source material rather than model memory.
- Validation-first — designed around GxP expectations: traceability, human-in-the-loop control, versioning and audit trails, so output is defensible to inspectors.
- Source-grounded — answers and drafts are anchored to an embedded library of regulatory guidance (e.g. EMA GVP modules) to minimise hallucination.
- Human authority preserved — agents accelerate; qualified people review, decide and sign off.
Status: In active build (greenfield).
Regulatory source library ─► retrieval / grounding
│
agent orchestration (task-specific PV/regulatory agents)
│
human-in-the-loop review ─► versioned, audit-trailed deliverables
│
React workspace · API · Postgres (Prisma)
TypeScript (monorepo, Turborepo) · Node.js · React · PostgreSQL · Prisma · LLM agents (Google Gemini + others) · retrieval-augmented grounding · GxP-oriented audit/versioning design · ESLint / Prettier.
Built by Dominic Gonsalves · open to contract / fractional AI roles