Convert SRT to Premiere Pro markers
This turns each SRT cue into a Premiere sequence marker, landing on an exact frame at the timebase you choose. Markers are not captions: they are navigation. If you want a transcript you can search and jump around in while you cut, this is the format that does it.
Drop a .srt file here
or click to choose one. It is converted in this page and never uploaded.
Premiere marker XML addresses frames, not seconds. This has to match the timeline you are importing into, or every caption lands slightly off and drifts further as the video runs.
Markers, not captions
Premiere reads SRT perfectly well as a caption track, and if you want captions on screen you should import the SRT directly. Markers do a different job: they put the transcript into the Markers panel, where every line is searchable and double-clicking one moves the playhead to the moment it was said.
For interview and documentary work that is the more useful artifact of the two. Finding the take where someone said a particular sentence stops being a scrubbing exercise.
Each marker carries the line twice
The Markers panel's name column truncates, so a long caption shows as a fragment. Each marker here has the shortened line as its name and the complete line as its comment, because marker search matches the comment. A line too long to display is still a line you can find.
Premiere cannot read FCPXML
This is FCP7 XML (xmeml), which is a different format from the FCPXML that Final Cut Pro 10 and DaVinci Resolve read, despite the similar name and shared ancestry. Handing Premiere an .fcpxml produces an error rather than a timeline. If you are working in Final Cut or Resolve, use the FCPXML converter instead.
In Premiere: File → Import, select the .xml, and open the sequence it creates. The markers are on the sequence itself, so you can copy them onto the sequence you are actually cutting.
A subtitle file records when each line appears, not when each word is spoken. Anything built from one is accurate to the line. If you need per-word timing (for karaoke-style captions, or to split a line without the second half drifting), transcribe the audio instead, which times every word individually.
Nothing is uploaded
The .srt is read and the .xml is written in this page, by the same code that builds exports in the editor. The file never reaches a server and no copy is kept. A subtitle file is a transcript, and a transcript is often the part of a project that cannot leave the building.
Why it asks for a frame rate
A .srt stores times in milliseconds and says nothing about frames. Premiere marker XML addresses frames directly, so something has to decide which frame each cue lands on, and only you know the timeline you are importing into. Pick the sequence's rate rather than the source video's: they are not always the same, and that mismatch is the usual cause of captions that start slightly off and get worse.
Frequently asked questions
Why does the frame rate matter for markers?
A marker's in and out points are integer frame counts at the sequence timebase, not timecodes. If the timebase here does not match your sequence, every marker lands at the wrong frame by a proportion that grows through the video.
Will this give me captions on screen?
No. Markers are navigation aids and never render. Import the SRT directly as a caption track if you want visible captions; you can do both from the same file.
Why is the marker name shorter than my caption?
The Markers panel truncates its name column, so the name is shortened to fit and the full line is kept in the comment. Premiere's marker search reads the comment, so you can still find any word in the line.
Can I import this into Final Cut Pro?
Not usefully. Final Cut reads FCPXML rather than xmeml. Use the SRT to FCPXML converter for Final Cut and DaVinci Resolve.
No subtitle file yet?
Generate one from your audio or video with word-level timing, then export Premiere marker XML directly, no conversion step, and every word individually timed rather than every line. Free for 15 minutes and 5 transcriptions a month.
Transcribe a file instead