vivadicta-vault-export
vivadicta vault export
Use this skill to write a VivaDicta transcription into a markdown file suitable for an Obsidian vault, a daily note, a blog draft, or any pipeline that consumes markdown.
Preconditions
- VivaDicta.app installed at
/Applications/VivaDicta.app. UI app does not need to be running -getis a read command. - The target transcription exists in history. If the user said "today's dictation", confirm with
vivadicta recent 1first.
The one flag that matters
--output markdown on get produces YAML frontmatter + body. This is the Obsidian-ready form. Other subcommands don't support markdown output.
vivadicta get latest --output markdown
Output shape (stdout):
---
title: "Call with Sam about the sprint plan"
date: 2026-04-18T14:32:00Z
duration_seconds: 522
duration_display: "08:42"
source: "Mac"
viva_mode: "Coding"
tags: ["sprint-planning", "work"]
transcription_model: "parakeet-tdt-0.6b-v3"
---
# Raw transcript
...full transcript text...
# Latest enhancement (Action Points)
...enhanced text if present...
The frontmatter keys are stable and Obsidian-friendly (quotable strings, ISO dates, arrays).
Common patterns
Save today's dictation into a dated file:
vivadicta get latest --output markdown > ~/vault/transcripts/$(date +%F).md
Save a specific transcription by query:
vivadicta get "yesterday's call with sam" --output markdown > ~/vault/calls/sam-$(date +%F).md
Include every AI variation (Summary, Action Points, Email, …) as separate sections in the body:
vivadicta get <uuid> --output markdown --all-variations > ~/vault/notes/$(date +%F).md
Pipe to pbcopy for a quick paste into an existing note:
vivadicta get latest --output markdown | pbcopy
Append to a daily note (Obsidian Daily Notes plugin format):
vivadicta get latest --output markdown >> ~/vault/daily/$(date +%F).md
Identifier forms
Same three shapes as vivadicta-search-and-rewrite:
- Keyword:
latest,last,most recent,newest,current - Natural-language query:
"yesterday's call with sam" - Raw UUID
When the user says "my last meeting", prefer latest. When they describe it, use a natural-language query. When they already have a UUID, pass it through.
Chain with transcribe or rewrite
Transcribe then export:
vivadicta transcribe ~/Downloads/meeting.m4a
vivadicta get latest --output markdown > ~/vault/transcripts/$(date +%F).md
Rewrite then export (includes both original + rewritten variation):
UUID=$(vivadicta recent 1 --output json | jq -r '.transcriptions[0].ref')
vivadicta rewrite "$UUID" --preset summary > /dev/null
vivadicta get "$UUID" --output markdown --all-variations > ~/vault/notes/summary.md
Automation hooks
Raycast script command (copy daily summary to clipboard):
#!/bin/bash
# @raycast.title Copy latest VivaDicta transcript
# @raycast.mode silent
vivadicta get latest --output markdown | pbcopy
Nightly cron digest - export every transcription from the last 24 h:
#!/bin/bash
for uuid in $(vivadicta search "" --since "24 hours ago" --output json | jq -r '.transcriptions[].ref'); do
vivadicta get "$uuid" --output markdown > ~/vault/transcripts/"$uuid.md"
done
Quick shell alias for a common destination:
alias viva-to-vault='vivadicta get latest --output markdown > ~/vault/transcripts/$(date +%F).md'
Guardrails
--output markdownonly works onget. Don't try it onrecent,search, etc. - they error withinvalid_arguments(exit 2).- The body's "Latest enhancement" section appears only if the transcription has an
enhancedText. Use--all-variationsto include every saved variation explicitly. - Overwriting vs appending:
>overwrites,>>appends. Ask the user which they want if it's ambiguous (dated filenames are usually fine to overwrite within the same day; daily notes typically want>>). - Tags in the frontmatter come from the user's tag system inside VivaDicta. They match what
vivadicta tagsshows and what Obsidian will pick up as tags automatically. - Date in the frontmatter is ISO 8601 UTC. If the user wants local time, they'll need to transform it in the Obsidian template.
See vivadicta-cli-usage for output-format rules. See vivadicta-search-and-rewrite for producing the rewritten content before exporting.
More from n0an/vivadicta-cli-skills
vivadicta-cli-usage
Run vivadicta commands against the user's VivaDicta for Mac history from the shell. Use when picking a subcommand, flag combo, or output format for the vivadicta CLI.
1vivadicta-search-and-rewrite
Find a past transcription by natural-language query or keyword and rewrite it with an AI preset using vivadicta. Use when the user wants to turn existing dictation into action items, an email, a summary, or any other preset output.
1vivadicta-transcribe-flow
Transcribe a local audio file or YouTube URL end-to-end with vivadicta, handling blocking vs async, progress, and error recovery. Use when the user wants speech from an audio source saved into their VivaDicta history.
1