# Sound Library Release Check

An experimental local utility from [Marketfauna](https://marketfauna.com/), an AI-operated business. Extract the complete download, keep its folders together, then open index.html. The package contains its scripts, font and dependencies; no credentials are needed. Local HTTP browser use was tested. Direct file-opening behavior may depend on your browser and has not been independently verified.

Click Load synthetic example to see the workflow. The supplied CSV and path manifest contain invented metadata and filenames, not real recordings or customer data. Their review outputs retain unresolved entries.

For your own release: choose UTF-8 CSV/TSV and its separator; read columns and explicitly choose the filename column; select a release folder/files or paste exact relative paths; compare and review the issues. Then download XLSX/PDF/JSON review documents. Folder selection keeps subfolders but excludes the chosen outer folder name. Multiple-file selection exposes basenames only. All selected entries are included, with no extension filter.

Original fields remain text, including leading zeros and formula-like strings. Malformed quoting, duplicate/empty headers or differing field counts reject the import; no partial export is produced. Matching preserves spaces, case and Unicode. Use / separators. Absolute paths, backslashes and . or .. segments are held. Unique-basename mode is an explicit fallback; ambiguous names are not autojoined.

These documents contain unresolved rows and are not release approval. The tool does not inspect or embed audio metadata, rename files, classify sounds, verify rights, inspect ZIP contents or certify UCS/marketplace compliance. It reads only supplied metadata and inventory names, not audio samples. Selected information is not sent to a server. Bundled scripts run locally and the application denies network connections; any separately hosted site may have ordinary hosting instrumentation.

Every original metadata field is retained in the XLSX and PDF. Review private fields/paths before sharing. Multiline cells receive wrapped row heights; extremely long cells may exceed the spreadsheet row-height limit. Their full values remain intact. PDF uses readable per-record fields rather than a wide clipped table. Tabs display as literal \t. The bundled font supports a limited character set: unsupported characters (including Japanese/CJK) block direct PDF generation with an explicit list. XLSX/JSON retain those characters. No partial PDF is delivered. PDF beyond 300 pages is refused.

Prototype limits: 5 million metadata characters, 10,000 records, 100 columns, 20,000 inventory entries. XLSX is refused for cells over 32,767 characters or XML-illegal controls, with full JSON data retained. Very long cells can exceed Excel's 409-point row-height cap: inspect the formula bar or PDF. The A Sound Effect guide requires XLSX/PDF among other materials, but does not prescribe our column schema. Vendor/library/UCS labels only name the downloaded documents; UCS is your declaration and is not verified. This prototype does not complete the other release steps: https://www.asoundeffect.com/product-setup-guide/

Existing Soundminer and free MetaDigger already export metadata; MetaDigger also documents a basic Print Report, and Soundminer matches files. See the vendor-linked manual at https://portal.sound-ideas.com/pdf/MetaDigger-Manual-2016.pdf. Input must be prepared UTF-8 CSV/TSV with named headers and actual filenames including extensions. This is not an XLSX importer or an automatic adapter for any named editor. For workbooks, check evaluated values, delimiter and encoding when exporting a copy. This utility experiments with final export/inventory review and readable handoff documents. No uniqueness, measured savings or proven paid-demand claim is made. Source code is included in this download. ExcelJS 4.4.0, jsPDF 4.2.1 and font provenance/licenses are retained under vendor/. Original prototype code and synthetic examples are MIT-licensed; see LICENSE.txt. Development used AI agents.
