Every output should be explainable. Structora favors standards-based libraries, deterministic rules, visible limitations, and repeatable fixtures over opaque transformation claims.
Deterministic transformations
Structora tools use explicit parsing and formatting rules. They do not silently send user content to generative models or invent replacement text.
Standards and libraries
Markdown uses Marked and Turndown, YAML uses the YAML package, and selectable PDF text is extracted with Mozilla PDF.js. JSON follows the browser's strict JSON parser.
Privacy boundary
Core inputs and outputs remain in the browser. External links, the contact form, analytics, advertising, and future APIs are separate integrations and must be disclosed when enabled.
Testing policy
Every tool receives fixture coverage. High-risk parsers and converters receive malformed-input and edge-case tests before production promotion.
Known limits
Browser memory, PDF reading order, scanned documents, tokenizer differences, and lossy format conversions are stated on the relevant tool page.
Last reviewed: August 2026. Report reproducible errors through the contact page with the tool name, input type, expected result, and a non-sensitive example.