Products

One engine. Three ways to consume it.

The same deliberative, grounded engine — packaged for how much you want to run yourself. Start free and local, move to a hosted team workspace, or have us deliver the finished study.

01The ladder

From self-serve to done-for-you.

01

Open Core

Local · free

Install the MCP server on your own AI agent and run unlimited councils on your machine. Your host model writes the text — we don’t sell tokens.

  • One MCP config block — set up in under a minute
  • No API key for the core; runs on your own agent
  • Unlimited councils & syntheses, stored locally
  • Synthetic-only & local-first — no PII to ingest
02

Cloud

Hosted · from €39/seat·mo

The same inspector UI, hosted and shared. Auth, multi-tenancy and billing are managed; you get team workspaces without running anything yourself.

  • Same UI as Open Core, plus the Cloud workspace
  • Shared accounts, roles & team workspaces
  • Managed auth, multi-tenancy & billing
  • Server-side rendered — no proprietary logic shipped to the browser
03

Research

Managed · from €4,500

Send the brief to research@sonaloop.com. We scope, run and quality-control the study, then deliver a PDF plus a portal you can connect over MCP. Fixed-scope, ~2 weeks, synthetic-only.

  • Email research@sonaloop.com — scoped and quoted within a day
  • We run it: councils, syntheses, human-in-the-loop QC
  • Your design system applied to the deliverable
  • Delivered as a PDF + a portal (MCP or ask-on-study)

The same engine underneath.

Every tier runs the open-core engine: a moderated council built to disagree, personas with memory, and an auditable synthesis. The difference is only how much you operate.

See how it works