Find any conversion in the /conversions directory

The homepage and /convert are where you do the work, but /conversions is where you look before you commit. It is the authoritative, always-current directory of every active input-to-output pair the current release can create and validate. Because the image side alone runs to hundreds of routes, no memorized or printed list stays accurate for long, so the directory is the single place that does not drift out of date.

Open convert.novusstreamsolutions.com/conversions and browse by input format. Structured-data routes for JSON, CSV, TSV, and TXT and archive routes for ZIP, TAR, and TAR.GZ live here alongside the image pairs. If a pair appears in the directory, the current release can genuinely produce and validate it. If it does not appear, treat it as not available in this release rather than assuming it is hidden somewhere.

An active route is more than a page that names two formats. Three things have to be present for a pair to be a working converter: a real upload control that accepts the input, an output selector that offers the destination, and a validation gate that checks the finished result before the download button appears. When all three exist for an input-to-output pair, that pair genuinely converts.

The validation gate is the part people overlook. A result must match its expected signature, container, or decodability before Novus Convert exposes its download, so a file can finish processing and still fail that final check. That is deliberate. Refusing a download that does not pass is safer than offering one that only has the right extension, and it is why an active route can be trusted rather than merely started.

Everything behind an active route runs in browser memory, including HEIC and HEIF decoding through a lazy-loaded local WebAssembly worker. Being active does not mean the file is sent away to a remote service; it means the conversion can be produced and verified on your own device.

The format directory is broader than the active converter, and that gap is exactly where mistakes happen. Pages for video, office documents, ebooks, CAD, and fonts exist as reference-only guides today. They describe a file type; they do not process it. A page does not become a working converter simply because it exists and names a format.

The reliable tell is the upload and output control, not the page title. On an active route you can add a file and pick a validated output. On a reference-only guide you can read about the format, but there is no upload workflow to begin. When you are unsure which kind of page you are on, look for the control rather than reading the heading as a promise.

Decide your input format and your destination format, then confirm that specific pair in the directory before you queue anything. A source you can upload does not guarantee every output, because the outputs available depend on the route. Checking the pair first saves you from loading files for a destination this release does not produce.

Think in concrete pairs rather than in loose format families. HEIC to JPG, PNG to ICO for a favicon, TIFF to PNG, JSON to CSV, and ZIP to TAR.GZ are each a distinct route with its own entry. The image outputs the directory can list for a source include JPG, PNG, WebP, GIF, APNG, BMP, TIFF, TGA, and ICO, and which of them apply depends on the input, so read the pair you actually need rather than assuming symmetry.

Once you add a file, Novus Convert identifies it by its signature, the actual bytes, rather than by the filename extension. That is why the outputs offered for a row can be trusted: they reflect what the file really is. A JPG renamed to .png is caught here, and the row shows the outputs for its true type instead of the label someone typed.

This is what connects the directory to the queue. The directory tells you which pairs are possible in general, and the signature check tells you which outputs are compatible with the specific file in front of you. When a file is mislabeled or malformed, that mismatch surfaces at this point rather than producing a broken result later. If the outputs you expected do not appear on a row, the file is usually not the format its name claims.

If a pair is not in the directory, the current release does not convert it, and the accurate reading is that the conversion is not available yet rather than tucked away somewhere. For the reference-only formats, video, office documents, ebooks, CAD, and fonts, the page exists to explain the format, not to process it, so there is no upload control to find no matter how long you look.

When your route does exist, remember the boundaries that come with local processing. Active image routes accept files up to 100 MB and 80 megapixels, structured text up to 25 MB, and archives up to 200 MB compressed with no more than 5,000 entries and 512 MB expanded. Daily download allowances are tracked per input format and reset by local calendar day, with the remaining count shown beside the download control, so free is not the same as unlimited.

Results live behind temporary object URLs in the current session. Removing a job or closing the page revokes them, so save what you need before you leave. The directory tells you a route exists; these limits tell you what a single session of it will hold.