Glossary

The words we use, defined once

Sapilon has a precise vocabulary because governance needs one. Here is what each term means, everywhere it appears.

Governance

Ownership zones: Explicit markers in the codebase that define what the AI may touch and what stays under human control. Three zones exist: //!AUTO, //!SAFE, and //!DEV. The zones live in the code itself, not in a hidden policy. More →
//!AUTO: The zone the AI owns: it may change this code autonomously, with every action logged, diffable, and reversible. More →
//!SAFE: The zone where the AI proposes and a human approves: changes are staged as diffs for review before they land. More →
//!DEV: The humans-only zone: the AI does not modify this code. More →
Governed agent: Sapilon’s AI agent, constrained by ownership zones and gates. Everything it does is logged, rendered as diffs, and reversible; automated code review runs on generated changes. More →
Expert verdict: The decision a verified human expert returns on an escalated action: approve, approve with changes, reject, or need more information. Reasoning is required for everything except a plain approve, and an approval clears the gate exactly once, for exactly the reviewed action. More →
Review brief: An AI-generated summary that carries everything an expert needs to decide: what was asked, what the agent was about to do, why it was blocked, and where the project stands. Most reviews are decided from the brief alone, with no project access. More →
Access grant: A scoped, time-limited permission under which an expert does hands-on work. It names the project and the expert, expires on its own, and can be revoked in one click. More →

Lifecycle

Project stages: The four gated phases every Sapilon project moves through: Design (architecture and frontend against mocked data), Build (backend wired up end-to-end), Rehearsal (pre-production dress rehearsal), and Live (real users, real data). More →
Build Server: The real development environment where your app runs end-to-end while you work on it. It starts when you need it and stops itself when you go quiet; each organisation’s first 50 hours are free, then €1 an hour. More →
Rehearsal Server: The pre-production AWS environment where user journeys are walked through, data migration is drilled, and security and costs are checked before go-live. On the Essential architecture the Live Server rehearses first instead. More →
Live Server: The production environment. The new system takes over here and the old one is retired; versioning, monitoring, and backups keep it healthy long-term. More →
Bootstrap: The automated first pass of a modernization: Sapilon reads your legacy inputs (source, database schema, screens, documents) as the specification your old system already wrote, and rebuilds a working app from them on an AI-safe architecture in days, ~10× faster and ~10× cheaper than a traditional rewrite. Sapilon inspects your system first and estimates up front how much it can recreate. More →
Modernization: Turning a legacy system or an AI-built prototype into production-ready, AWS-native software you own: real architecture, tests, governance, and real cloud infrastructure. More →
Upgrade-aware diff: A computed diff that brings your app in line with the latest version of the Sapilon platform and its dependencies, so it stays current without a risky manual rewrite. Platform updates follow the same ownership zones as every other change. More →
Project memory: The architecture, intent, constraints, and decisions that persist with a project, so the AI keeps context instead of forgetting it between sessions. More →

Running your app

Platform: The open-source web and api foundation every Sapilon project is built on: sign-in, the UI kit, layouts, theming, the API contract, the database and its migrations. It goes public on 1 December 2026. More →
Architecture: How your app runs for real people, chosen from two questions about downtime and sensitive data: Essential (one server for everything), Balanced (a Rehearsal Server when you need one), or Dedicated (everything of its own, in an AWS account kept apart from other projects). More →
Checkup: The review a version passes before it reaches a Rehearsal or Live Server: setup, code (tests, types, security, migration safety) and protection. Findings are fixed, waived, queued as tasks or handed to an expert. More →
Built-in services: What the platform gives every app without new code: sign-in, users and roles, integrations and secrets, plus files, notifications and AI when the app needs them. More →

Commerce

Point: Sapilon’s fixed-price AI usage unit with one published rate. Every AI interaction costs a predictable whole number of points instead of raw token pricing. More →
Wallet: The prepaid balance everything draws from: AI points, AWS hosting, and expert time. Pay-as-you-go, no subscription, no seats, and the balance cannot go negative. More →
Feature Pack: A packaged, versioned feature slice (contacts, invoicing, bookings) extracted from a real Sapilon app, ready to apply to any project. More →
Gift card: Promotional credit from Sapilon, sent by email with a single-use code. Applying it adds the whole amount to your wallet; it has no cash value. More →
Job size: How expert work is priced: Small (under 30 minutes) and Medium (about an hour) are fixed prices; Large jobs are billed per hour against an estimate and a cap you approve before work starts. More →