AI infrastructure for clinical trial sites

Reverie — Turn site knowledge into faster study startup.

Reverie learns how your site reviews protocols, builds budgets, negotiates language, and launches studies — then applies that knowledge to every new trial.

Study startup is still done manually.

Site-side work runs on understaffed teams, startup included. This costs both sponsors and sites real money while delaying life-saving treatments.

180 days — Academic medical center activation median, against a 90-day target
$300K — Lost by sponsors for every week a study sits unactivated

AI has a memory problem

AI models don't know your site. Right now, there's no way for them to learn at scale.

Context windows lose relevant details
Limited knowledge of site history
Unreliable audit trails and citations

Meet Reverie

Start with how your site works.

Reverie integrates into a site's existing workflow and uses AI to process the site's history and practices into a navigable knowledge base.

Knowledge sources: previous studies, internal policies, budget history, approved language, standard operating procedures, and operational decisions.

Bring the protocol into that context.

When a new protocol arrives, Reverie maps it against the site's existing knowledge base and similar work from the past.

From the protocol: visits, procedures, labs, enrollment targets, timelines, and staffing requirements.

Procedure → site rate
Contract clause → approved language
Operational requirement → internal policy
Study task → prior study decision

Generate the work required to launch.

With the study and site together, Reverie helps teams create grounded documents, answer questions, assist with staff training, and more.

Outputs: study budget, contract edits, consent language, operational checklist, startup timeline, study tasks, and risks & exceptions.

Every completed study strengthens the site.

Reverie preserves which decisions were made and why. Each completed study improves the starting point for the next one.

One system across study startup.

Protocol Intelligence

Convert unstructured protocols into structured requirements, procedures, timelines, risks, and tasks.

Example protocol extract: visit schedule → 12 visits; labs → CBC, CMP, ECG; enrollment → 48 subjects.

Budget Automation

Connect study requirements with site rates, historical negotiations, assumptions, and prior decisions.

Example budget draft: screening visit, $1,240; infusion, site rate; pharmacy fee, prior study.

Document Intelligence

Review and generate study documents using approved language, internal policies, and site history.

Example document review: CTA §7.2, approved language; ICF risk language, matched; exception, flagged for review.

About us

Reverie was founded by Johns Hopkins researchers who experienced clinical trial site delays firsthand. We watched promising studies take months to activate due to manual tasks that relied on information fragmented across people and inboxes. We began building AI-native software for research sites to bridge that gap and help advances in clinical research reach patients faster.

Viraaj Reddi — Co-founder, Johns Hopkins, Biomedical Engineering and Computer Science
Atharva Mulay — Co-founder, Johns Hopkins, Chemical and Biomolecular Engineering
Milun Jain — Co-founder, Johns Hopkins, Biomedical Engineering

Turn study startup into a repeatable system.

Work with Reverie to configure the platform around your institution's documents, workflows, historical decisions, and study-startup processes.

Monthly platform subscription for research sites, site networks, and CROs; pilots available.

Built for clinical trial sites, site networks, and academic research organizations.