Behavioural research, on demand.
Reactor Self-Service puts CloudArmy’s platform directly in your hands. Configure a template study, field it to real respondents, and get standardised, decision-ready results — without waiting on a managed-service team.Your platform. Your studies. Your timeline.
Reactor Self-Service is the fully online, client-operated way to use CloudArmy’s platform — template-based studies you configure, field, and analyse yourself, with results delivered automatically.
Start with a question
Every study starts with a business question — here: which pack design wins with US shoppers, and why?
The Reactor app
Reactor is the online app where you set up, launch, monitor, and manage every study — from your browser, on your schedule. Self-service is direct access to run your studies yourself.
Template-based studies
Start from proven study templates — like tournament testing to narrow a large set of options fast. You configure and field it; no bespoke setup required. Log in to browse the full template library and join the waitlist for the suites you want released first.
Automated analysis
The workflow runs end to end and delivers a standardised output package — a full array of charts, graphs, and images ready to drop straight into your presentations and deliverables, with no bespoke analyst intervention.
Nothing proprietary ever reaches a model
Reactor MCP strips every piece of proprietary information — PII, product names, customer names — and replaces it with pseudonyms before anything reaches a model. The agents only ever see the pseudonyms; you see your real names, remapped automatically on the way back.
PanelGuard data quality
Behavioural-science data-quality checks identify fraud, disengagement, and low-quality responses — so you trust what you measure. Included on every plan.
Zero-touch delivery
Access the tool through your account, run the study, and receive results — all without manual operational involvement from CloudArmy.
Reactor StudioDive deeper into insights
Reactor Studio is the desktop analysis tool that goes beyond the automated report — the same Analysis Studio CloudArmy’s own analysts use, now in your hands. Included with Professional and Enterprise.
Vibe — or hand-write — R and Python through MCP
Integrated AI assistant on every top model
Create and save repeatable analysis & packaging — images and outputs tailored to your organisation
Charts, plots, and deep-dive exploration
Supercharge your science with Reactor MCP
Reactor MCP puts 150+ models at your fingertips — always on their latest versions, updated as the labs ship. Frontier models for reasoning and analysis, plus vision-and-behaviour models built for stimulus, packaging, and creative work. And because every prompt is sanitised through Reactor MCP, nothing proprietary ever reaches a model.
Frontier modelsTwo ways to buy.
Go all-inclusive with a monthly subscription, or pay à la carte — a low platform fee plus per-test pricing, with recruiting billed per panelist. All prices in USD. Log in to join the waitlist for the plan you want released first.
Pause or cancel anytime.
Cancel anytime — you keep what you’ve paid for through the current month, then it simply ends. No future charges. Or pause instead: we hold your place and waive the usual start-up fee when you come back, so pausing beats cancelling.
Starter
$3,000
/ month · USD
For small insights teams or individual researchers exploring Reactor.
Reactor app access
1 template suite
PanelGuard included
Automated fielding
Automated reporting
Email support
Professional
$7,500
/ month · USD
For established insights teams running regular research on Reactor.
Reactor app access
Up to 3 template suites
PanelGuard included
Automated fielding
Automated reporting
Reactor Studio (Analysis Studio) with AI credits
Priority support
Enterprise
$10,000
/ month · USD
For organisations that need benchmarking and the full Reactor Studio toolset.
Reactor app access
Up to 5 template suites
PanelGuard included
Automated fielding
Automated reporting
Reactor Studio with expanded AI credits
Benchmark access & creation
Benchmark database creation
Benchmark model generation
Real people. Hard science. Citations, not hallucinations.
Every study runs with real, authenticated human respondents — never AI imagining what people might do. Agents do the heavy lifting of analysis and packaging, but every conclusion is tied back to your actual data, with citations and the evidence behind it. We don’t make claims — we show you the receipts. AI accelerates the analysis; it never invents the findings.
Real, authenticated humans
Studies field to verified human respondents, and PanelGuard weeds out fraud and disengagement — so the data reflects genuine behaviour.
Agents do the work
AI agents handle the heavy lifting of analysis and packaging — fast and repeatable, always grounded in your actual data.
Hard science, cited
Findings come from proven behavioural-science methods, with citations and the underlying evidence behind every claim — no black boxes.