For institutions
Universities, hospitals and statistics offices get SPSS-class analysis and an AI orchestrator without sending a single record to someone else's cloud.
Who it's for
Each field answers to someone different: a supervisor, a reviewer, an ethics board, a minister or a client. OrcStat is built for all of them.
Guided test selection, plain-language help with a four-part structure for every procedure, and APA 7 output that drops straight into a thesis.
Mixed models, repeated measures, SEM, mediation and reliability, with a syntax journal that makes every analysis reproducible for reviewers.
Survival analysis, ROC, Bland-Altman and complex survey data on patient-level records that never leave the hospital network.
Complex samples, custom tables and large datasets with audit trails, offline operation and admin-enforced policy.
Custom tables with significance letters, multiple response sets, weighting, conjoint and AI-assisted coding of open-ended answers.
Privacy and security
Every control works offline, is set by policy, and leaves an audit trail.
Security and data handlingAI requests are pinned to approved hosts. DNS, redirects, IPv6 and IPv4-mapped addresses are re-checked, and blocked attempts are logged.
Projects choose how much an AI model may see, from variable metadata only to full data, with a preview of every request.
AES-256-GCM project files with scrypt-derived keys, encrypted autosave and recovery, and SPSS-compatible password-protected .sav.
API keys and passwords are stored by the operating system, never in project files or logs. Passwords are redacted from the journal.
Each AI request records what was shared and what was approved. Export it for governance, or purge it on schedule.
Extension packages are hash-checked and Ed25519-signed. An admin policy can refuse unsigned packages entirely.
Deployment
Silent installs, a single admin policy file and offline operation. A licence check sends only the licence ID, a hashed device ID, the app version and the operating system, never your data. Air-gapped machines can use a permanent offline activation instead.
Ask about deploymentSigned macOS app with a sandboxed App Store build. Windows and Linux are next on the roadmap.
The full engine runs in the browser through WebAssembly after users sign in. Nothing is uploaded, which suits locked-down teaching labs.
Point every seat at a shared on-premises model server, or let each user run a local model.
Headless job runs from the command line, with launchd or cron schedules for production reporting.
# One file for every seat ai: local_only: true endpoint: http://llm.internal:8000/v1 endpoint_locked: true max_exposure_level: 1 # aggregates only extensions: require_signed: true telemetry: off
Admin policy
Lock the model endpoint, cap how much data the AI may see, require signed extensions and switch off telemetry for every install.
Pilot programme
We're working with a small number of universities and research groups during the beta.