
How to Edit Training Videos
Training videos have to be followed, not just watched. A new employee pauses, rewinds and copies each step. They may be watching on a shop floor with no sound, in a second language, or with a screen reader. Software training adds one more need: the screen has to stay readable while the trainer stays on screen. This guide covers procedures, onboarding, safety and software training for teams.
The short answer
Cut the recording so each step of the procedure is its own short section with a title, show the screen or the hands doing the task while the trainer explains it, and add captions. For software steps, keep the screen recording full frame and put the trainer's webcam in a small box in a corner, so learners can read every click and still see who is talking. Record the screen with test data, or hide customer details before you import it. Then send it for approval, and make accessible and translated versions for staff who need them. Eddie builds the steps, titles, webcam inset, captions, audio description and review link from prompts.
The deliverables
What you will end up with
- A video where each step is a titled section that is easy to find again.
- The task on screen while the trainer explains it.
- On the software steps, the screen full frame with the trainer in a corner.
- Captions, and a screen recording that shows test data, not customer details.
- An audio-described version and a translated version for the staff who need them.
6 steps
Step by step
- 01Import the recording and the screen capture
- 02Cut it into titled steps
- 03Show the task, keep customer data off the screen
- 04Put the trainer in a corner over the screen
- 05Add captions and get approval
- 06Make accessible and translated versions
Step 1
Import the recording and the screen capture
Upload the trainer's recording, the screen recording of the software and any footage of the physical task. Eddie transcribes the trainer and logs the screen recording as b-roll. If the screen recording carries the trainer's voice and lands as interview footage, ask Eddie to treat it as b-roll before you build the edit: the trainer's camera stays the main footage, and that is what the webcam inset in step 4 is made from.
What Eddie does
- Transcribes the trainer, so each step can be found by what is said.
- Logs the screen recording and the hands-on footage by what is on screen.
What you get: Everything the training needs, matched by what is said and shown.

Step 2
Cut it into titled steps
Ask for one section per step, each with a title. Learners come back to a training video to find one step; titles make that possible.
What you type “Edit the returns training into 6 steps, one section each, following the order in the trainer's recording. Put a numbered title at the start of each step, such as "Step 1, Check the order number". Natural pacing, with a short pause after each instruction.”
What Eddie does
- Builds the edit in six sections from the trainer's soundbites, at the pacing you asked for.
- Places a numbered title at the start of each section.
What you get: A procedure that reads like a checklist.

Step 3
Show the task, keep customer data off the screen
Put the screen recording over the software steps and the hands-on footage over the packing steps. Deal with customer data before the screen recording reaches Eddie. Eddie hides areas on the main footage only, not on a b-roll clip laid on top of it, and here the screen recording is b-roll. So record the software with a test account and test orders, or cover the name and address panel with a solid box before you import. Use a solid fill, not a blur, for names and addresses: a few recovered letters can be enough to read them.
What you type “Show the returns system screen recording during steps 1, 3 and 4, and the packing station footage during step 5. Keep the trainer's voice running underneath.”
What Eddie does
- Places your screen and hands-on footage over the matching steps, with the trainer's voice running underneath.
- Hides areas on the main footage only, the trainer's camera, never on the screen recording laid over it, so that recording has to arrive clean.
What you get: Learners see each action on a screen recording made with test data or cleaned before import.

Step 4
Put the trainer in a corner over the screen
On a software step, learners need the whole screen to follow the clicks, and they still want to see who is talking. Keep the screen recording full frame and show the trainer's own camera in a small box in a corner while the screen is up. When the screen goes away, the trainer is back full frame. Leave the box off the hands-on footage, where it only gets in the way. The layout has room for one inset at a time.
What you type “While the returns system screen recording is on screen, show the trainer's own camera as a small box in the bottom right corner, about a quarter of the frame wide. Leave it off the packing footage in step 5.”
What Eddie does
- Lays the trainer's own picture, muted and at the same moment, on a second video layer in the corner you named. It is picture-in-picture from one camera, with no multicam group needed.
- Shows the box only while a full-frame cutaway covers the trainer, so the same person is never on screen twice, and drops it on the segments you name.
- Can crop the box to the trainer's head and shoulders, or cut the trainer out of the background: a free cut-out, or a cleaner paid one that Eddie quotes first.
What you get: The clicks stay readable and the trainer stays in view on every software step.

Step 5
Add captions and get approval
Captions are essential on a noisy floor. Then send a review link to the process owner and compliance so they approve the exact wording.
What you type “Add minimal captions at the bottom, then give me a review link so the process owner and compliance can comment at the timecode.”
What Eddie does
- Adds captions from the word timings.
- Creates a review page that needs no Eddie account, and collects comments on the timeline.
What you get: An approved, captioned training video.

Step 6
Make accessible and translated versions
Staff who are blind or have low vision need audio description; staff who work in another language need a dubbed version. Make both from the approved edit.
What you type “Make an audio-described version of the approved training for staff with low vision, then export the approved version and the audio-described version as MP4s.”
What Eddie does
- Writes and voices a description wherever the picture carries information the narration does not, as a new edit beside the original.
- Renders both versions. Caption files such as SRT download from the web app's Export menu.
What you get: One training, usable by everyone on the team.

Craft
What makes it work
- One step, one section
- Learners rewind to a step, not to a minute. Titled sections make the video work as a reference.
- Show hands and screens
- Narration alone does not teach a procedure. Film the hands doing it and capture the screen.
- Open full frame, then go to the corner
- Start with the trainer full frame for the introduction, then move them to the corner when the screen comes up. Camtasia's own webcam tutorial offers the same two layouts, full-screen camera and picture-in-picture, and suggests the camera can fill the screen at the start of the video.
- Doing the inset by hand in Premiere
- Put the webcam clip on a video track above the screen recording: Premiere builds the picture from the lowest track up, so clips on upper tracks cover the ones below. Select the webcam clip and choose Window > Effect Controls. Expand Motion, lower Scale, then drag the clip into the corner in the Program Monitor, or type the Position values in Effect Controls.
- Record with test data
- Set up a test account and test orders before you capture the screen. Eddie cannot hide part of a b-roll clip, and a clean recording needs no hiding at all. If real data does get captured, cover it with a solid fill before import and before any review link goes out: a blur can leave text readable.
- Plan for every learner
- Captions, audio description and translation are cheaper to make from the approved edit than to retrofit later.
FAQ
Questions
- Can Eddie put my webcam in the corner of a screen recording?
- Yes. Keep the trainer's camera as the main footage, lay the screen recording over it as b-roll, and ask for the trainer in a corner. Eddie shows the box only while the screen covers the trainer, so the same person is never on screen twice, and you can leave it off any segment you name.
- Do I need a multicam group for a webcam inset?
- No. The inset is the trainer's own recording, muted and shown at the same moment as the main picture, so it stays in step with the voice without syncing a second camera.
- Can Eddie make training videos accessible?
- Yes. Eddie adds captions, and it can make an audio-described version of an edit for blind and low-vision viewers, with a still-frame pause wherever a description needs more time.
- Can I translate a training video?
- Eddie can dub a finished edit into another language, placing the translated lines where the originals land. See the dubbing page for details.
- Can I update one step later?
- Re-record the step, add it to the project and ask Eddie for a new edit that uses the new take for that step. Eddie builds it as a new edit beside the approved one, so send the new version for review before it replaces the old video.
- Is it safe for internal content?
- Eddie does not use your files, projects, prompts, transcripts or edits to train AI models. Single sign-on is for enterprise customers: the Eddie team sets it up, so contact sales first; you cannot turn it on yourself. Eddie cannot hide part of a screen recording laid over the trainer, so record with test data or hide customer details before import.

Edit your next training video with Eddie
Sign up, add your footage and type the first prompt above. New accounts start with free credits.