A template-driven ingestion and normalization method for converting arbitrary human-facing inputs into deterministic, machine-readable structured data.
Version: 1.0
Status: Stable Draft
scanOS is a method and product for transforming screenshots, photos, scans, documents, or text fragments into structured representations using predefined templates and deterministic normalization rules.
It provides a consistent, predictable way to extract visible information and convert it into formats suitable for downstream processing, analytics, and automation workflows.
scanOS is implementation-neutral:
it is not tied to any specific programming language, OCR engine, runtime,
or AI model. Any system that follows the method is considered an implementation
of scanOS.
- Template-driven ingestion
- Deterministic normalization rules
- Transparent handling of uncertainty
- Structured output (JSON, YAML, XML, CSV, etc.)
- Compatible with batch, real-time, and interactive workflows
- Designed for pipeline integration with any downstream system
(including MetaMemoryWorks modules such as trainingOS, nutritionOS, sleepOS)
- Not an OCR model
- Not a hallucination-proof system
- Not a prediction or reasoning engine
- Not guaranteed to extract data perfectly
- Not tied to a single implementation or technology stack
scanOS defines how ingestion and normalization must be performed, not the specific tools used to perform it.
The full technical description of the scanOS method is available in:
This document defines:
- inputs
- processing pipeline
- template system
- normalization rules
- error model
- implementation neutrality
scanOS is free for individual, non-commercial use.
Corporate, institutional, or governmental use requires a commercial license.
The full licensing terms are defined in:
→ scanOS_RESTRICTED_LICENSE_v1.0.md
The license also defines:
- derivative works
- restrictions on commercial use
- prohibitions
- warranty disclaimer
The method, structure, and conceptual origins of scanOS are documented in:
This establishes authorship, chronology, and prevents patent capture.
scanOS may be implemented:
- in any programming language
- with any OCR backend
- with or without language models
- on local or cloud systems
- in CLI tools, APIs, or applications
Format-shifting, media-shifting, or implementation-shifting does not remove a system from being considered an implementation of scanOS if the method is followed.
scanOS uses semantic versioning for the specification.
Templates may version independently.
For licensing inquiries or questions regarding commercial use, please contact:
Johannes Glaser
(Author and Copyright Holder)