Coaching review is a short loop: cut the rep out of a long session, pull stills for a sequence sheet, slow the clip down, then send a version the athlete can open on a phone. Every step runs in the browser, so a client's training footage never leaves the device.
The order matters. Trim before slowing, cropping or compressing, because re-encoding a three-second rep is far faster than re-encoding a twelve-minute session. Extract frames from the trimmed clip rather than the full video, and remember that slowing footage does not add frames: a 30 fps clip at 0.25x holds the same 30 frames per second of source, shown longer. Real slow motion starts in the camera, at 120 or 240 fps. Work on a copy so the original session file stays intact for the next review. Nothing here needs an account, and no step uploads a file.
1. Cut the rep out of the session
Video Trimmer puts the clip on a timeline with two handles. Drag the start and end points, or play the video and press Set start here and Set end here. Precise mode is the default and re-encodes the selection so the cut lands on the frame you picked. Fast mode copies the streams without re-encoding, so it is nearly instant and leaves the quality untouched, but it can only begin on a keyframe, which means the clip may start a second or two early.
For coaching, a ten to twenty second window around the rep is usually enough, and a shorter clip makes every later step faster. One range per export, so two separate moments mean two passes. If the video comes from a phone in HEVC and the timeline stays empty, convert it to MP4 first, because a container the browser cannot decode will not load here at all.
Video Trimmer
Mark the rep on the timeline and export it in fast copy or frame-accurate mode.
2. Pull frames into a sequence sheet
Extract Frames has three modes. Single mode saves one frame at a time you type or scrub to, accurate to a hundredth of a second. Interval mode takes one frame every N seconds across the whole clip, and count mode spreads a fixed number of frames evenly from the first moment to the last. Both batch modes are capped at 30 frames per run.
For a movement breakdown, interval mode is the one to use: a 0.2 second interval on a three-second rep gives 15 stills, enough to see the start, the contact and the finish. PNG keeps every pixel the video holds, while JPG and WebP are smaller and have a quality slider. A scale-down option is there for web sheets, and it never enlarges a frame that is already smaller than the target. Frames are decoded one at a time, so a full batch takes a moment.
One thing to be clear about: the tool reads frames that exist. A 60 fps phone clip gives 60 chances per second, and a 24 fps camera gives 24. Extraction cannot invent an in-between position that was never recorded.
Extract Frames
Save up to 30 stills per run at a fixed interval, in PNG, JPG or WebP.
3. Slow the clip down for review
Video Speed Changer runs from 0.25x to 4x in 0.25 steps, with presets for 0.5x, 1.5x, 2x and 4x. The 0.25x end is reached by dragging the slider. Audio defaults to Keep pitch, which stretches the soundtrack through an atempo chain so a voice stays natural instead of sounding speeded up; Mute audio drops the track entirely, which is often what you want for a pure visual check with your own commentary over it.
The output is MP4 at CRF 20, and the panel shows the new duration before you start. Two honest limits. Slowing does not interpolate frames: at 0.25x each existing frame is simply held for four times as long, so a fast limb still jumps between positions. And 1x is refused on purpose, because re-encoding a clip to change nothing wastes time and quality. If the athlete films at 120 fps or 240 fps, the slow motion will look far smoother than anything produced from a 30 fps file.
Video Speed Changer
Slow to 0.25x with the voice kept natural, or mute the clip for a silent review.
4. Crop a vertical version for a phone
Most athletes watch on a phone held upright, and a wide gym shot wastes half the screen. Video Cropper sets the frame to 9:16 and lets you choose whether the crop sits at the top, the middle or the bottom of the source, which is how you keep the athlete in frame when the camera was pointed high. The output height can be 1080p, 720p or 480p, or a custom size such as 1080 by 1920 when the platform wants full vertical resolution.
Cropping throws away pixels on the sides, so frame the shot in the camera when you can rather than relying on the crop. Cropping also re-encodes the clip, which is a second generation of compression after the trim and the slow-motion pass, so do it once and at the size you actually need. For a quick check on a phone, 720p is plenty and encodes much faster than 1080p.
Video Cropper
Crop to 9:16, choose the position, then export at 1080p, 720p or a custom size.
5. Join angles or several athletes
Video Merger takes two or more clips, lets you drag them into order, and joins them into one file. Clips can be resized to a shared 720p or 480p height, or kept at the size of the first clip, so a mix of phone footage and camera footage ends up consistent. Use it to put a front view and a side view of the same lift one after the other, or to send a coach a single file with three athletes instead of three attachments.
This is a sequence, not a comparison layout: there is no picture-in-picture, no split screen and no overlay. Joining also re-encodes the clips, so the file goes through one more generation of compression. Keep the individual clips if you may need them again.
Video Merger
Join two or more clips in order and normalise them to a shared frame size.
6. Compress before you send it
Chat apps and messaging links have limits that a 4K slow-motion export will hit immediately. Video Compressor exposes quality, resolution, frame rate and output format. For a coaching clip, 720p at 30 fps is usually the sweet spot: the movement is still readable on a phone, and the file drops by 50 to 80 percent compared with the original. Lower the quality slider until the file size fits the channel you are using.
Compression is one-way, so keep the trimmed original. There is no target size field, which means you judge by the output number rather than typing a limit. Clips under 200 MB and a few minutes long work best, because the whole encode happens in browser memory; a very long session should be trimmed first rather than compressed in one pass.
Video Compressor
Set quality, resolution, frame rate and format, with live progress and no upload.
What this chain cannot do
- No frame interpolation. Slow motion repeats existing frames. A 30 fps clip at 0.25x is not a 120 fps recording, and the difference is visible on a fast limb.
- 30 frames per extraction run. Interval and count modes stop at 30 stills; single mode has no batch cap but saves one frame at a time.
- No drawing or annotation. Nothing here draws an arrow, an angle, a line or a circle on the frame, and there is no picture-in-picture.
- No frame stepping. The preview seeks in time, not frame by frame, and it does not show a frame number.
- No automatic movement analysis. There is no pose tracking, rep counting, speed measurement or comparison against a model.
- Every export is a re-encode. Trim, slow, crop, merge and compress each cost a generation of compression, except fast-mode trim, which copies streams.
- No team library or share link. Files are produced on your device and downloaded; sending them is up to you.
- No proprietary camera formats. If the browser cannot decode the file, convert it in the camera maker's app first.