Everything user-visible, newest first, including the bugs we caused ourselves. Missing a tool? Request it.
Every tool used to hide its options once it produced a result, so changing one setting meant uploading the file again. The controls now stay put: extract six colours then ask for ten, convert to WebP then to JPG, move a watermark to another corner - same file, no re-upload.
A site security header blocked camera and microphone access outright, so the browser never showed a permission prompt and recording silently failed. Access is now requested properly, and when something is genuinely blocked or in use by another app the recorder explains which and what to do about it.
The screen recorder can now show your webcam in a corner of the screen recording, in a single combined video. Pick the corner and the size, and change either one mid-recording without stopping.
The image watermark tool now stamps your own logo, not just text. Upload a PNG, scale it against the width of the photo, and place it in any corner at the opacity you want.
The unit converter now covers area, speed, time, data, pressure, energy, power, angle and frequency alongside length, weight, volume and temperature, with more units in the original four. The free API endpoint gained the same categories and still accepts the unit names it always did.
Voices are now grouped by language with a dedicated language picker, starting on your browser's own language. Install a system voice and the language appears in the list.
The homepage now sets out plainly how FileOven differs from the typical online converter, including where the alternatives are the better choice, and links to the full comparison for each one.
Every AI tool is marked coming soon rather than left to fail part-way through a run. They will return once a new provider is wired up.
Every major format now has a reference page explaining what it is, what it cannot do, and which format to use instead, linked to the tools that handle it. Added guides on converter privacy and choosing a web image format.
Tools are now browsable by category, and every tool page links to related tools and the reverse conversion where one exists.
Tool pages previously rendered their text only after the interactive widget loaded. The description, steps, specifications and questions now arrive with the page itself, so they are readable instantly and on slow connections.
Each tool now states plainly whether it runs in your browser or on our servers, whether files are uploaded, what a run costs, and whether it works offline.
The speech transcription tool displayed the price of a run as roughly $0.10 when the correct figure is $0.50. The calculation is now derived from the published credit price so it cannot drift again.
Loading a tool with the embed parameter served the full site layout instead of the embed layout, and password recovery links could fail to open the reset form. Both now work as documented.