Short videos

Correct the words, style the captions and export the video in the same editor. No separate subtitle app.

Try first: Import a short speaking clip, add it to the timeline and turn on autocaptions. Check the words before exporting.

A personal Microsoft account is required. Your audio is processed online. Free projects using your own or free assets export without a watermark; 4K and premium assets cost extra. You can also download an SRT subtitle file.

Upload the recording you already have. Enhance Speech reduces background noise without rebuilding the edit.

Try first: Open Enhance Speech and upload a short voice recording. Listen to the before and after before keeping the result.

The free plan allows one audio file at a time, up to 30 minutes and 500 MB, with 1 hour total per day. Video uploads and strength controls are paid features. Keep the original recording in case you prefer it.

Edit the transcript to remove a spoken mistake. You can make a first cut without hunting through every second of the timeline.

Try first: Import a short talking video. Correct its transcript, delete one unwanted sentence and play across the cut.

Free processing is capped at 60 media minutes a month. Watermark-free video export is paid, and AI tools have their own credit limits. Listen across edits: a clean-looking transcript can still create an awkward cut.

Start with suggested moments, reframing and captions. Useful for checking whether your interview has a good Short hidden inside it.

Try first: Upload a video you own and review one suggested clip. Check that it makes sense without the rest of the conversation.

The free plan has a watermark, no editing, and disables export after 3 days. You need a paid plan to remove those restrictions. Its clip or virality scores are suggestions, not proof that a post will perform well.

Try different voices and pacing on the same paragraph. You can audition a voiceover, not just read a script silently.

Try first: Paste a short script into Text to Speech and choose a library voice. Listen for names, numbers and unnatural pauses.

The free allowance is limited and does not include a commercial license. Paid Starter lists commercial licensing. Use a licensed library voice or a voice you have permission to use; do not clone someone without consent.

Balance speech, music and different speakers in an existing recording. It tackles uneven volume, not just microphone noise.

Try first: Upload a short recording with uneven voices, enable the leveler and compare quiet and loud sections after processing.

Free processing adds an Auphonic jingle. Paid credits remove it, and unused free monthly hours do not roll over. Keep your original: automatic leveling and cleanup may change the sound in ways you dislike.

Keep each person's audio and video separate so you can fix one guest's level or framing without changing everyone else.

Try first: Create a studio and make a short test with a consenting guest. Let their upload finish, then inspect the individual tracks.

The free separate-track allowance covers your first 2 recorded hours, not 2 new hours every month. More separate-track downloads require an upgrade. Editor exports can include video or audio branding; inspect the export options before recording a full episode.

Upload a recording in your browser and download an SRT file. A simpler starting point than installing command-line transcription software.

Try first: Upload a short, non-sensitive clip, choose its spoken language and transcribe it. Check names and timing, then export SRT.

Requires signup and uploads the recording to a cloud service. Free files run one at a time with lower priority. Speech recognition still needs checking. An SRT is a separate subtitle file; you need an editor to permanently put the captions onto the video.

Use local Whisper transcription through a desktop interface when sending a recording to a cloud service is not a good fit.

Try first: Check the official project's installation notes for your computer. If the distribution fits your security requirements, try a short sample with a local model and review its transcript.

Not a phone or one-tap browser tool. Models need downloading and speed depends on your computer. The project warns its Windows installer is unsigned; do not bypass security warnings just to try it. Check current platform support and review transcription errors before using the SRT or VTT export.

Explore how lyrics or a musical mood might sound. Useful for an early creative sketch, not as a promise of free music you can put in a published video.

Try first: Describe one mood and instrument, or use lyrics you wrote yourself. Listen to a draft to decide whether the idea is worth developing.

The free plan lists 50 credits per day, a shared queue, no monthly song downloads and no commercial rights. Paid download allowances are also capped. Check the terms for your particular track before publishing, especially if it uses uploaded or remixed material.

Turn a script into a stock-avatar explainer when the missing piece is someone presenting it on screen. This is different from a voiceover over still images.

Try first: Choose a stock avatar and a short script you wrote. Generate one simple explanation, then check the pronunciation and visual result before using another free slot.

Watermark removal is paid. Premium avatar generation and lip-sync translation have separate trial or credit limits; do not assume the free plan covers those workflows. Use authorised material and make clear when a presenter is AI-generated where viewers could otherwise be misled.

Useful when you need an audio file to put into an editor, not just a voice preview you cannot export. Start with a short sentence to check the voice and pronunciation.

Try first: Paste two sentences you wrote, choose a language and voice, then generate and listen. Download the audio promptly if the pronunciation works for your project.

Individual conversions have shorter limits, and some voices have different allowances. Generated files are removed after 30 minutes. The provider permits commercial use of generated audio, but that does not clear rights to someone else's script or other material.