Building The Executive Study Bible required embedding 3,074 verse objects — across two translations, 66 books, 1,189 chapters — into a local JavaScript data array. There was no tool that could do this reliably. AI chat interfaces cannot execute thousands of sequential API requests without context degradation, token limit collapse, and hallucination risk. Doing it by hand was not viable. A dedicated extraction tool had to be built before the main product could exist.
That tool was Version 1 — a purpose-built API compiler, functional and austere. It worked. And having built it, a question emerged that most builders would not think to ask: what if the tool itself became extraordinary?
Version 2 arrived. Then Version 3 — The Luna Rosa SoT. Schema of Things. A cinematic command center for any paginated JSON API extraction.
The name carries intent. Luna — the moon, the orbital, the thing that circles the thing that matters. Rosa — the rose, the living, the thing that grows toward light. SoT — Schema of Things — a declaration that what is being built here is not just a data tool, but a framework for understanding the structure of information itself.
Building The Executive Study Bible required 3,074 verse objects across two translations. AI chat interfaces cannot execute thousands of sequential API requests without context degradation and hallucination risk. A dedicated extraction tool was required before the main product could exist.
Beyond the Bible project: the broader problem is that browser-native, zero-installation, configurable JSON API extraction tools do not exist as purpose-built products. Enterprise data extraction tools require infrastructure, subscriptions, and technical expertise. Developer tools are code-first and inaccessible to non-programmers. No cinematic, configurable, single-file extraction environment existed that could be handed to a non-developer and returned structured data.
The Luna Rosa SoT was not built to compete with enterprise ETL pipelines. It was built for the builder who needs to extract structured data from a paginated API — accurately, locally, without hallucination risk, and without installing anything.
Three progressively capable extraction tools. The Luna Rosa SoT — Version 3 — runs any paginated JSON API extraction inside a cinematic environment: a live Three.js globe at 23.5° axial tilt, a procedural parallax cityscape across three SVG layers, a laser strike sequence on extraction complete, a holographic terminal logging every action, and a cursor hijack that signals the moment the job is done. The operator configures the API endpoint, sets the index range, defines the transformation rule, and watches the machine run.
Before the Bible app could exist, Barb the Builder had to write out thousands of Bible verses in a very specific format that a computer could read. Instead of doing it by hand — which would have taken weeks — she built this tool to do it automatically: accurately, quickly, and at scale. It does the boring, careful, essential work so the beautiful work can begin. Version 3 wraps that work inside a visual environment that makes the machine feel as significant as the task it is performing.
Sequential API loop, configurable start/end index. Dot-notation path traversal walks parsed JSON to target array. User transformation rule via Function constructor with live syntax error detection and fallback. Three.js WebGLRenderer, 128-segment sphere, 23.5° axial tilt, requestAnimationFrame eased spin. Laser: three feGaussianBlur layers, feComposite merge, SVG filter chain. Procedural cityscape: three SVG layers generated on load and resize. Cursor hijack on extraction complete: cursor:none, floating div tracks mousemove, click event restores default cursor. Terminal logs every step as it happens.
The global web scraping and data extraction market was valued at $754 million in 2024 and is projected to reach $2.87 billion by 2034 at a 14.3% CAGR (Market.us, 2024). The broader data extraction market — spanning web, document, database, and API sources — is valued at $6.16 billion in 2025 and headed for nearly $28.5 billion by 2035 (Market Research Future, 2025). Sixty-five percent of enterprises now use automated extraction to feed machine learning pipelines (Mordor Intelligence, 2024).
The market is growing rapidly — but it is bifurcated between enterprise managed services requiring significant infrastructure, and developer-first open-source frameworks requiring code expertise. The gap between them — a configurable, browser-native, zero-installation extraction tool accessible to a non-developer — is the position The Luna Rosa SoT occupies. Not competing with Bright Data or Apify for enterprise scale. Occupying the space where the builder who needs to extract 3,074 structured objects from a public API — accurately, locally, without hallucination risk — has nowhere else to go.
The data extraction market is also experiencing a specific pressure point: AI-generated hallucination in sequential extraction workflows. Market Research Future notes that AI-driven web scraping improved accuracy to 99.5% in controlled conditions — but AI chat interfaces were not designed for multi-thousand sequential API request loops with strict schema requirements. The Luna Rosa SoT was built precisely because no AI interface could be trusted with the task it was designed to perform. That is not a market positioning statement. It is the documented origin of the product.
The three-version evolution — purpose-built compiler to general instrument to cinematic command center — documents how this studio thinks when it encounters a constraint: not by accepting it, but by building through it. A pure CSS engineering showcase dressed as a data infrastructure tool. The globe spins at 23.5° axial tilt because the Earth does. The laser strikes because the moment of extraction completion deserves a signal. The cursor hijack is not a flourish — it is the machine telling the operator: the job is done. Your hands are yours again.
No user data collected. API calls run client-side during extraction only — no proxy server, no intermediate logging, no third party that ever touches the data in transit. Function constructor sandboxed in browser. No backend. No database. No telemetry. On extraction complete, the cursor is hijacked — cursor:none, a floating div tracks mousemove — signaling the operator that the job is done. A click restores the default cursor, and the extracted output is available in the terminal to copy and use. Nothing is transmitted or retained by the tool itself after the session ends.
Think of the Luna Rosa SoT as a very private courier. When it extracts data, it makes the API call directly from the operator's browser — no middleman, no server in between that could intercept or log what passes through. The data lands in the browser session. The tool holds nothing, stores nothing, and sends nothing to anyone. That behavior — no unauthorized access, no data stored externally, no analytics capturing what passes through — is what HIPAA calls Technical Safeguards. The tool itself is not the liability. It does not discriminate by industry, and neither does its privacy architecture. Healthcare, legal, financial, education, government, research — the safeguard holds across all of them.
There is one boundary worth naming clearly: the Luna Rosa SoT connects to whatever API the operator points it at. It has no control over what that API does on its end. If the source is sitting on an unsecured server, logging every request, or owned by a party with no data protection policy — the Luna Rosa SoT cannot fix that. The courier's conduct is clean. The warehouse it picks up from is the operator's responsibility to vet.
No documentation site required — the CRT terminal logs every action as it happens. The machine teaches itself to the operator in real time. Every version of this tool was built under constraint. The constraint is what made each version necessary. That is not a workflow. It is a philosophy.