cri-one
Alchemy Data Three

Alchemy Data Three

$12B
$12,000,000,000
SKU 2070

Go to cart

Why you'd want it

The third version of a compound-analysis research system, for teams who want the research engine.

  • Who it's for: chemistry and materials-science research groups, and buyers of research software IP.
  • The problem it solves: keeps the compound-ranking engine and its research notes together in one package.
  • What you get: reference Python code and research notes, on encrypted Blu-ray.

Asset valuation: $2,400,000,000. Master index of all projects: PROJECTSINDEX. Last edited: 2026-05-02 Status: Third-generation Alchemy Data substrate + supporting inventor records Maturity (per PACKAGEARCHITECTURE.md): L1 (Research / Working Software) — research code, blog imports, and inventor rec

Payment is by bank transfer: through Stripe up to $999,999.99, and by wire invoice above that. Questions: christopher@cri-one.com.
Contents 16 files in 1 folder · listed 14 November 2026
  • .claude/
    • settings.local.json 1 KB
  • 1 light trigger.txt 742 KB
  • abnormal.odf 7 KB
  • alchemy-blog-posts.wxr.xml 20 KB
  • alchemy_data_three_example.py 1 KB
  • CHANGELOG.md 1 KB
  • CITATION.cff 1 KB
  • commons.odf 7 KB
  • Contact Information.txt 1 KB
  • CONTACT_INFO.txt 1 KB
  • LICENSE 1 KB
  • MANIFEST.json 2 KB
  • PROPOSED_DESCRIPTION_EDIT__2026-10-21.md 1 KB
  • README.md 3 KB
  • reallight.odf 6 KB
  • unique.odf 6 KB

Description

Alchemy Data Three — Third-Generation Compound Analysis System

Alchemy Data Three is the third generation of Christopher Gabriel Brown's compound-mechanism analysis system — the computational substrate that powers the disease-discovery research behind the Alzheimer's, Parkinson's, and 300-disease-database patent applications. It builds on Project 07 (Alchemy Probability Data V2), wrapping that engine in a new layer; a nonprovisional patent filing gives the system its legal backbone. A buyer who acquires this package gets the machine that produces the discoveries , not just one disease's worth of output.

How v3 wraps the v2 engine

Alchemy Data Three is not a rewrite. It is a superset. The directory tree mirrors the entire Alchemy Probability Data project (Project 07, the v2 substrate) inside a 07-alchemy-probability-data/ subtree, then adds new third-generation content on top. At the code level, the v2 engine's AlchemyDataV2 class imports and wraps three subsystems — the ElementTransformationDatabase , the FormulaSynthesisMapper , and the original v1 probability libraries ( ProbabilityCalculator , MultiplierCalculator , MultiDegreeProbabilityEngine ) — into a single compound-discovery pipeline. V3 preserves this full architecture, with the nonprovisional patent filing establishing the legal filing for the combined inventions and connecting the substrate to the disease-discovery applications.

Package at a glance — 5,576 files
Extension Count What they are

.gz 5,210 Gzip-compressed JSON data snapshots (probability matrices, compound chunks, transformation records)

.svg 195 Scalable vector figures — compound diagrams, patent drawings, pipeline schematics
.csv 41 Ranked compound tables scored across multiple dimensions

.md 28 Markdown documentation (quick-start guides, system architecture, test results, changelogs)

.py 18 Python scripts — the analysis engine, API server, data generators, diagram renderers

.json 14 Index manifests, configuration files, recipe records, integration mappings
.txt 13 Plain-text documentation, inventor notes

.html 11 Web interfaces (compound search, synthesis explorer, transformation browser, valuation pages)

.odf 8 Working ODF research notes ( abnormal.odf , commons.odf , unique.odf , reallight.odf )

(no ext.) 7 Licence files, extensionless markers
.bat 4 Windows batch launchers (web interface, data generator, progress check, CLI)

.zip 2 Compressed archive bundles (AlchemyValueIndex v3, ScientificCalculator)

.ts 3 TypeScript type declarations for code-signing modules
.js 3 JavaScript code-signing modules

other ~17 .cff citation, .xml blog export, .ps1 disc-burn script, .pyc bytecode cache, .map source maps

The 18 Python scripts

Every script lives inside the blueprints/ directory of the mirrored v2 engine. Together they form the complete analysis, generation, and access stack.

Three additional .py files (totalling 18) are present in the assembled package's data-access, core-systems, or utility handoff directories; they provide the original v1 probability library ( alchemy_probability_library ), the multi-degree probability engine ( alchemy_multi_degree_probability ), and supporting helpers that alchemy_data_v2.py imports.

The 41 CSV data tables

The CSV tables are the distilled, human-readable output of the analysis engine. Each row represents one element combination scored across nine dimensions (see "Probability matrices" below). The tables are organised into four compound categories — metals (alloys and intermetallics), solvents (organic, protic, and ammoniacal), gases (single-phase mixtures and noble-gas blends), and functional objects (exotic multi-element compounds) — and four scoring slices.

By-category, by-slice breakdown (19 core tables)
Slice Metals Solvents Gases Objects

Clean (safe + productive, no hazardous flags) clean_metals_top10.csv — clean_gass_top10.csv clean_objects_top10.csv

Nice (nice-and-productive, balanced value) nice_metals_top10.csv nice_solvents_top10.csv nice_gass_top10.csv nice_objects_top10.csv

Novel (rarity-favouring, may include exotic/hazardous) novel_metals_top10.csv novel_solvents_top10.csv novel_gass_top10.csv novel_objects_top10.csv

Safe (no radioactive, toxic, or reactive flags) safe_metals_top10.csv safe_solvents_top10.csv safe_gass_top10.csv safe_objects_top10.csv

Full-spectrum value tables (4 tables)

Four additional CSVs — value_metals.csv , value_solvents.csv , value_gass.csv , value_objects.csv — rank every combination by composite value score, segmented into top, middle, and bottom slices. Each row carries the full nine-column property vector. For example, value_metals.csv shows that AlFeSi (ternary intermetallic, value 17.66, Mohs 7.0, $1/kg, availability 9.75) leads the metals, while quaternary intermetallics like AlFeMgSi (value 17.15) and FeMgSi (value 16.82) follow.

The remaining 18 CSVs in the full package are generated outputs, including the 96-333-fixed.csv master table and additional per-chunk exports that the find_novel_outputs.py script produces during full data sweeps.

The 2,609-file data corpus

The bluray_data/ directory is the heart of the system — a self-contained probability database designed to fill a 25 GB Blu-ray disc. It contains 2,609 files (2,605 gzip-compressed JSON archives plus four plain-text/JSON metadata files).

Corpus structure

195 SVG figures

The SVG collection breaks down into seven functional categories.

The remaining SVGs in the full 195-file count are additional generated compound diagrams and patent-figure variants produced during extended data sweeps.

The 4 blog posts

A WordPress export file ( alchemy-blog-posts.wxr.xml , WXR format, 20 KB) contains four long-form articles intended as companion documentation for the system:

The "41-41-41" subtree — what it adds over v2

The 41-41-41/ folder is the third-generation layer that transforms the v2 data engine into a legally filed invention. The nonprovisional application text, its USPTO forms and the filing receipt are kept by the inventor and are not part of the package.

In short, v2 is the engine; the 41-41-41 subtree is the patent filing that turns the engine into filed intellectual property.

Connection to disease-discovery patents

The Alchemy Data substrate underlies three disease-discovery patent applications in the inventor's portfolio:

All three applications depend on the compound-discovery pipeline, the probability matrices, and the transformation database that ship with this package. Acquiring Alchemy Data Three means acquiring the substrate that powers these filings.

Compound-mechanism analysis methodology
The system's analysis pipeline operates in five stages:

Probability matrices — scoring dimensions and weights

Every compound combination in the system is scored across nine dimensions, each computed from per-element property tables built into the find_novel_outputs.py script:

Transformation databases — synthesis procedure structure

The ElementTransformationDatabase stores transformations as structured records. Each transformation has:

The FormulaSynthesisMapper extends this with complete invention recipes: given a formula, it outputs named synthesis methods with temperature, equipment, multi-step instructions, and success probabilities. The system currently maps nine physics/chemistry formulas (E = mc², E = hf, PV = nRT, F = ma, and others) to concrete synthesis routes. The method_catalog_analyzer enumerates 257,500+ base known methods (772,500 with condition variations) across categories: enzymatic reactions (100,000), named organic reactions (50,000), catalytic reactions (30,000), inorganic reactions (20,000), materials synthesis (15,000), polymerisation (10,000), electrochemical (8,000), and more.

Data archive format — .gz snapshots

The 5,210 .gz files are gzip-compressed JSON documents. Each preserves a complete, self-describing data snapshot:

All archives use Python's gzip and json modules for read/write. The BluRayDataReader class handles transparent decompression and caching when querying the corpus.

What v3 adds over Project 07 (Alchemy Probability Data V2)

Web, Python, and CLI access paths
Web interface

Run python bluray_data_api.py (or double-click START_WEB_INTERFACE.bat ) to start the Flask REST API on localhost. Then open bluray_data_web.html in a browser. The tabbed interface provides compound search, transformation lookup, synthesis method retrieval, compound discovery, invention recipe generation, and dataset statistics. Additional web pages include formula_synthesis_web.html (formula-to-synthesis explorer with success probability display) and how_to_transformation_web.html (transformation browser).

Python API
from bluray_data_reader import BluRayDataReader

reader = BluRayDataReader("bluray_data") results = reader.search_compounds(['H', 'O'], max_results=10) transformations = reader.find_transformations('C') methods = reader.get_synthesis_methods('H2O') discovery = reader.discover_compound(['H', 'O']) recipe = reader.generate_invention_recipe(['H', 'O']) The BluRayDataReader class handles gzip decompression, caching (configurable cache size, default 1,000 entries), and index lookups transparently. For deeper analysis, import the core engine directly: from alchemy_data_v2 import AlchemyDataV2 .

Interactive CLI

Run USE_DATA.bat or python bluray_data_reader.py interactive to enter a menu-driven command-line session. The menu offers compound search, transformation lookup, synthesis methods, compound discovery, recipe generation, and statistics display — the same capabilities as the web interface, without a browser.

Batch and disc-burning

RUN_BLURAY_GENERATOR.bat launches large-scale data generation. CHECK_PROGRESS.bat monitors the run. CREATE_BURNABLE_DISCS.ps1 (PowerShell) prepares the generated data for Blu-ray burning. These tools target 25–50 GB datasets for archival distribution.

Supporting materials

How this is delivered

This package ships on a single Blu-ray disc, posted to the delivery address on your order. The archive on the disc is AES-256 encrypted; the passphrase is sent separately, by email, once the disc is despatched — so the disc alone is of no use to anyone who intercepts it.

Discs are despatched by tracked, signed-for post. Allow 5 business days for mastering and verification before despatch.

Footnotes

  1. Product number 8489.
  2. Family: alchemy data.
  3. Purchase: When you buy, you buy the use of that blueprint only.

For the Curious

Think of a chemist's notebook that never stops asking "what happens if I put these two together under heat, pressure or a catalyst" - and then answers itself, scores the answer and files it. Alchemy Data Three is the third generation of that notebook. It is the computational substrate underneath the disease-discovery research behind the Alzheimer's, Parkinson's and 300-disease-database patent applications, and it is research work: a machine that produces candidates for study, not a treatment and not a claim of one.

It is not a rewrite. The whole second-generation engine sits inside it, untouched, with a new layer wrapped around it - the probability calculators, the element-transformation database that records what carbon, hydrogen, nitrogen and oxygen do under heat, pressure, mixing and catalysis, and the formula mapper that turns known formulas such as PV = nRT and F = ma into step-by-step synthesis recipes. One script sweeps 36,068 unique element combinations and ranks them by novelty; another catalogues more than 257,500 base methods across nine categories. On top sits the 41-41-41 folder: a complete nonprovisional patent filing package that gives the system its legal backbone.

The package is 5,576 files - 5,210 compressed data snapshots, 195 vector figures, 41 ranked compound tables, 18 Python scripts and eleven web pages for searching compounds and browsing transformations. It is for a research group or licensee that wants the machine, not one disease's worth of output. The buyer walks away with the engine, the data and the filing record; the inventions stay credited to Christopher Gabriel Brown.