Exporting Transcripts
Export your transcript as PDF, subtitles (SRT/VTT), JSON, plain text, or your AI summary.
When a session is finished, you can take your transcript with you. Kalima's Export dialog turns the same transcript into a printable PDF report, ready-to-use SRT or VTT captions, a complete JSON file, plain text, or your AI summary as Markdown. Every format carries the speaker names you assigned.
How to export
Speaker names are pulled fresh every time the Export dialog opens, so your latest renames are always included. If you renamed someone, just reopen the dialog before exporting.
Choosing the right format
| Format | Best for | Notes |
|---|---|---|
| Sharing a record, archiving, client deliverables | Printable report with title, date, speakers, time ranges, original and translation text, plus the AI summary if one is saved | |
| SRT | Adding captions in a video editor | Timecodes use comma milliseconds; includes speaker prefixes |
| VTT | Captioning web-hosted video | Web-standard formatting for HTML5 players |
| JSON | Backups, your own processing, feeding another tool | Full structured data including word-level timing |
| Summary (Markdown) | Pasting outcomes into notes or a wiki | Requires a saved AI summary first |
PDF report
Choosing PDF opens a styled, printable report in a new window. It includes the session title, an export date, speaker labels, time ranges, the original text and any translation, and your AI summary if one has been saved. From the new window, use your browser's print dialog to print it or save it as a PDF.
Only finalized text is included in the PDF. Preliminary words still being processed are left out so your report stays clean.
Subtitle exports (SRT and VTT)
Both subtitle formats produce timed captions with speaker prefixes. SRT (SubRip) is the common choice for desktop video editors; VTT (WebVTT) is built for web video players and accessibility on websites.
Choosing the original or translation track
When your session includes translations, a track dropdown appears in the Export dialog. Pick Original to caption the spoken language, or Translation to produce captions in the translated language. This lets you generate captions for an international audience from the same session.
Caption timings are estimated from the recording's word timing, with a minimum cue length, so they are close but approximate. Nudge cues in your editor if needed, or use the JSON export for precise token-level timing. Only finalized segments are included in SRT and VTT files.
JSON export
JSON gives you the complete picture: every segment with its text, translation, speaker, detected language(s), and word-level timing data, plus session details and word and character counts. It's the right choice when you want a full backup or plan to feed the transcript into another tool, a spreadsheet, or your own workflow.
AI summary export (Markdown)
If you've generated and saved an AI summary for the session, you can download it as a Markdown file containing the overview, key points, action items, topics, decisions, and open questions in one portable document.
The Summary (Markdown) option stays disabled until an AI summary has been generated and saved for that session. Generate the summary first, then it becomes selectable.
Speaker names in exports
Every export carries the speaker names you assigned. Names are loaded fresh when the dialog opens, so they always reflect your most recent edits. Any speaker you haven't given a custom name keeps its default "Speaker N" label.
Frequently asked questions
Five: a printable PDF report, SRT captions, VTT captions, a complete JSON file, and your AI summary as Markdown. SRT and VTT let you choose the original or translation track when the session has translations.
Names are loaded fresh when the Export dialog opens, so close and reopen the dialog to pick up your latest renames. Any speaker you haven't given a custom name keeps its "Speaker N" label by design.
It only becomes available once an AI summary has been generated and saved for that session. Generate the summary first, then it turns selectable.
They're estimated from the recording's word timing, with a minimum cue length, so they're close but approximate. If you need precise data, the JSON export contains the underlying token-level timing.
The track dropdown only appears when the session has a translation. With no translation there's a single track to export, so the dropdown is hidden. To get a translation track, the session must have been transcribed with translation enabled.
Both are timed caption files. SRT (SubRip) is widely used in video editors; VTT (WebVTT) is the format web video players and HTML5 pages expect. Choose SRT for editing software and VTT for web playback.