How to Import SRT Into Premiere Pro From a Transcript Your Client Checked
The Premiere steps to import an SRT file, plus the part every tutorial skips: getting the client to proof the words before a single caption lands on your timeline.
Premiere Pro captions on a course module my agency delivered went out already burned into the picture, and we hopped on the review call feeling pretty good about it, right, because the edit was tight and the color matched the brand deck. About ninety seconds in, the client paused the video and said, very politely, that the name of their own framework was spelled wrong. It was spelled wrong every single time it came up across a twelve minute lesson, so we fixed the caption track, exported and uploaded all over again, all for something the client could have caught in two minutes by reading the words first.
To be very honest, the Premiere side of that fix was the easy part, because once you know how to import SRT into Premiere Pro the caption track basically builds itself, right. The real problem was the order we did things in. These days in my agency the client proofs the transcript before a single caption touches the timeline, and only then does the editor pull the SRT into Premiere, style it and send the captioned cut back as the next version.
So this post is that whole loop, the Premiere steps plus the proofing around them that nobody writes about, which at the end of the day is where the hours actually go.
Why import SRT into Premiere Pro instead of typing captions
Typing captions by hand is fine for a thirty second ad, right, but an interview, a course lesson or a podcast episode runs to hundreds of lines, and every one needs a start time, an end time, the correct words and so on. Adobe Premiere Pro has had its own speech to text for a few versions now, and I'm pretty sure most of you have tried it. It does a decent job, but that transcript lives inside your project, so the only person proofing those words is the editor, and the editor is usually the person who least knows how the client spells their own product names.
An SRT file is really really simple when you open it in a text editor. Each cue is a number on its own line, then a start time and an end time joined by a little arrow and written as hours, minutes, seconds and milliseconds, like 00:04:12,480, then one or two lines of caption text, then a blank line before the next cue begins. Because the timing is in milliseconds and not frames, the same file works on a 24, 25 or 30 frames per second sequence, and it can also go to YouTube as a sidecar or to a course platform, right. I wrote about the course side of this in reviewing captions on course videos, and the short version is that one clean SRT saves you from fixing the same typo in four places.
Getting a transcript your client can proof before it becomes captions
The first thing I do is upload the locked cut to PlayPause and run the AI transcription on it. That gives you a clickable transcript sitting next to the player, so when the client clicks a line the video jumps to that moment and they can hear it and read it at the same time. The client opens one link in their browser, with no account and no install, and they can proof the whole thing in roughly the time it takes to watch the video once, which is honestly all the time most clients will give you anyway.
When they spot a wrong word, they leave a frame-accurate comment on that moment with the correct spelling, and if a whole sentence is off they drop a range comment across it instead. The catch here is that you have to tell the client what to look for, because if you just say "please check the captions" they will skim it, right. So for any SRT caption review I ask for four things specifically: names of people, brand and product names, numbers and prices, and industry jargon, since those are the words speech recognition trips on and the words the audience notices first.
A caption is just the transcript wearing a font, so if the words are wrong before import they stay wrong on every frame after it.
I also tell them not to worry about timing or line breaks, because that part is my job in Premiere, and asking a client to judge caption timing is a really really good way to get forty comments about nothing. If you want the bigger picture on how transcripts fit into a team, getting started with AI transcriptions walks through it, and for long interview work the searchable captions for documentary editors page shows how the same transcript helps you find soundbites.
AI transcription SRT export after the client signs off
Once the comments are in, I export the SRT from PlayPause, which comes with AI transcription on the Agency plan and up. I open the file in a plain text editor, TextEdit in plain text mode on a Mac or Notepad on Windows, and I use the client's comments as my checklist. For instance, a comment at 04:12 saying the surname is spelled with an "i" takes me straight to the cue that starts at 00:04:12. I fix the word and leave the cue number and both times exactly as they are, right, because one stray edit on a timing line can make Premiere skip that cue or refuse the whole file.
I fix the words in the SRT itself and not later inside Premiere, because the SRT is the file that goes everywhere else too, and if I only fix the caption track then the YouTube sidecar still carries the old typo. Podcast teams feel this one the most, since one episode turns into a pile of clips that all need the same spelling, which is why searchable transcripts for podcast producers matter so much on a weekly show. When you save, keep the encoding as UTF-8 so accented names and curly quotes survive. If you use Save As, check that the name still ends in .srt and not .srt.txt, because Premiere will not treat a text file as captions.
The other thing, and trust me on any level on this one, is that the transcript has to come from locked picture, right. If you trim two seconds out of the middle after the client proofed the transcript, every caption after that trim is now two seconds late and you end up retiming half a video by hand. So I lock the sequence before transcription the same way I would before a final review, which is the whole idea behind locking a Premiere Pro sequence in the first place.
Frame-accurate note, everyone sees the exact same thing.
How to import SRT into Premiere Pro step by step
Every tutorial covers this bit, so I will focus on what actually goes wrong in my agency. Before anything, check that your sequence frame rate matches the file you uploaded for review. Then check that the sequence starts at the same point the review file started, because if your review export had a slate or a few seconds of black at the head, the SRT timing counts from that head and your captions will land late by exactly that much.
- Open the sequence you exported for review and park the playhead at the very start.
- Go to File, then Import, pick the corrected SRT, and it lands in the Project panel as a captions item.
- Drag that item to the very start of the timeline above your video, and if Premiere asks for a caption format, choose Subtitle.
- Or skip the drag, open the Text panel, go to the Captions tab and choose the option to import captions from a file.
- Scrub the first caption, one in the middle and the last one to confirm they sit on the right words.
- If everything is offset by the same amount, select all captions on the track and nudge them together.
- Picture locked before transcription
- SRT saved as UTF-8 and still ending in .srt
- Sequence frame rate matches the review file
- Sequence head matches the review export
- First, middle and last captions spot checked
If the first caption is right but the last one is late, the SRT is almost never the culprit, right, because that pattern usually means the picture changed after transcription. At that point it is honestly faster to run the transcription again on the new cut than to retime three hundred captions. Does that make sense, right, a constant offset just needs a nudge, whereas an offset that keeps growing through the video needs a fresh transcript. And if the captions sit on the track but you cannot see them in the Program Monitor, check the visibility toggle on the caption track header before you assume the import failed.
Styling and timing captions on the Premiere caption track
Premiere Pro captions from SRT come in looking plain, and the trick is to style one caption and then save that look as a Track Style, so every caption on the track updates together instead of you clicking through them one by one. You do that in the Essential Graphics panel, or the Properties panel in newer versions, and once the Track Style exists you can reuse it on every episode for that client, which is basically a small brand kit for captions.
For readability I keep captions to two lines of roughly 32 to 42 characters each. I split long captions in the Text panel wherever the speaker breathes, and I merge tiny one word captions into the line before so they do not flash on screen. On vertical cuts I push the captions up above where the platform buttons sit and I check them on an actual phone, because what looks centered on a big monitor sits right under the like button on a real device, you see what I mean here.
When you export, the Captions section of the export settings lets you burn the captions into the video or create a separate sidecar file. For YouTube I usually upload a clean master plus the SRT, and I follow YouTube's recommended upload encoding settings for the master. If you are burning captions in for social, check the export against the review copy before it goes out, and my notes on Premiere Pro export colors looking different cover the most common reason a burned in export looks off.
Checking the captioned cut as a new version
This is the step I added after that course module call, and it keeps the whole loop honest. I upload the captioned export to PlayPause as the next version on the same card, so the client sees MV2 stacked on top of MV1 and nobody hunts through email for the latest file. It's the same idea as version control in software, right, one place where every version lives in order.
On Agency and up there's side-by-side version compare, so MV1 and MV2 can play next to each other and it becomes obvious that the only thing that changed is the captions. This last round of caption proofing before publish usually takes a client a few minutes, because they already approved the words on the transcript and they are really only checking placement and line breaks now.
The client finds typos after captions are burned in and you export and upload twice
The client flags wrong words on the transcript first, then checks the captioned cut as MV2
If you edit in Premiere, the PlayPause panel for Premiere Pro puts those comments inside Premiere, so you click a comment and the playhead jumps to that frame, and you can upload the next version straight from the timeline, which I covered in more depth in syncing client comments to Premiere. And if part of your team works in Resolve, my post on moving from Premiere to Resolve explains how review works there without a native panel, where you export the cut for review and work through the notes in the editor.
Frequently asked questions
Why won't Premiere Pro import my SRT file?
Most of the time the file is not really an SRT anymore, right. Check that the name ends in .srt and not .srt.txt, that it is saved as UTF-8, and that every cue has its number, its timing line and a blank line after it. If a timing line got edited by accident, compare that cue with a good one above it and fix the numbers, then import it again.
Why are my captions out of sync after import?
If every caption is off by the same amount, the sequence head probably differs from the review file, so select all captions and nudge them together. If the offset grows through the video, the picture changed after transcription, and transcribing the new cut again is faster than retiming hundreds of captions by hand, which is why I always lock picture first.
Which PlayPause plan do I need for the transcript and SRT export?
AI transcription and the SRT export are on the Agency plan and up, so Agency at $19 a month or Enterprise at $27 a month, both flat per workspace with members included. Creator at $9 a month gives you the frame-accurate comments and version stacks but not the transcription, and every plan comes with a 7-day free trial.
Can I burn the captions in and still deliver a separate SRT?
You can, and I do that on most social projects. You export one version with the captions burned in for platforms that need them baked in, and you keep the corrected SRT as a sidecar for YouTube and course platforms, and because you fixed the words in the SRT itself, both deliverables carry the same approved text.
If you want to run this loop on your next interview or course project, everything I used here lives on the Agency plan, so check the PlayPause pricing page and try it on a real client job during the 7-day free trial.
So yeah. That's my way of saying it.
Saumyajit co-founded PlayPause after years watching review and approval quietly eat creative teams' deadlines. He writes about the workflow side of video, feedback, versioning, and getting to a clean sign-off.
Related resources
Keep reading
Bring your team into one review space
Centralize feedback, lock approvals, and deliver faster, start free today.
Sign Up for Free