How the collection is organised
The searchable catalog contains 2,623 canonical tools. A tool is a named task with a specific input, operation and result. Related tasks can use a shared engine: converting a PNG image to WebP and converting a JPG image to WebP remain distinct input-format tasks, without implying two independent algorithms.
The site also contains reference and unit-pair pages, topic collections, guides and redirects. Those pages are not added to the capability count. The 107 topic hubs group related work; equivalent tag and category destinations are consolidated instead of creating multiple competing archives.
Local processing or server processing
Modern browser tools run their core logic on your device. Server tools send an input file or public media URL to a processing service. The processing filter in the directory distinguishes these modes where verified metadata is available. Original retained tools without that metadata should be checked individually.
Local processing can be limited by available memory and browser APIs. Server processing has upload, output, duration, concurrency and execution limits. A tool’s instructions and error messages take precedence over a general description of the collection.
Calculations and numerical precision
Calculators depend on the values and units you enter, the formula in use and any stated assumptions. Most numerical operations use JavaScript floating-point arithmetic; displayed results may be rounded. A larger number of displayed digits does not make an estimate more accurate.
Use included examples to check that you understand the model. Verify unit selection, time periods, percentage conventions, signs and whether fees or taxes are included. Financial, health and engineering estimates require additional context before being used for consequential decisions.
File conversion and document quality
Formats support different features. A conversion may flatten transparency, reduce quality, change metadata or omit unsupported layout features. Password-protected, damaged, unusually large or uncommon files may fail even when their extension appears supported.
Inspect the generated result in an appropriate application. For a document, check page count, legibility and layout; for an image, check dimensions and transparency; for audio or video, check playback, duration and synchronization. Keep your original until you have verified the output.
What testing can establish
The release process includes automated checks of registered tool examples, representative file operations, shared interface interactions, metadata, internal links and installation behaviour. Tests use particular fixtures and environments; passing them does not prove correctness for every possible input.
Browser differences, extensions, network restrictions and upstream platform changes can affect results. Live platform availability is separate from a local test of downloader code. The diagnostics page checks important capabilities on your own browser and installation.
Content, sources and corrections
Tool pages describe their purpose and supported controls. Where a calculation or technical definition depends on a reference, consult the listed source and confirm that it applies to your context. Search-demand estimates and tool counts are not promises of search traffic, indexing or rankings.
If an output is wrong, send the exact tool URL, a minimal non-sensitive input, selected options, expected result and actual result. An explanation of the expected result or a reference makes a report much easier to investigate. Use the report builder to prepare those details.