Skip to content

transcripts.md

Latest commit

 

History

History
214 lines (151 loc) · 6.17 KB

transcripts.md

File metadata and controls

214 lines (151 loc) · 6.17 KB

Transcript Commands

Query and search meeting transcripts from your Tana workspace.

Overview

Tana stores meeting transcripts as separate nodes linked to meeting entries. The supertag CLI provides dedicated commands to:

  • List meetings that have transcripts
  • Show full transcript content with speaker and timing info
  • Search within transcript text only

By default, transcripts are excluded from general search and embeddings to keep results clean. Use these dedicated commands when you specifically want to find spoken content.

Commands

List Meetings with Transcripts

supertag transcript list

Shows all meetings that have associated transcripts.

Default output (TSV for piping):

M9rkJkwuED    Monthly Team Standup    127    2025-12-20
xY7nKpQrSt    Q4 Planning Session     342    2025-12-18
aB3cDeFgHi    Client Review Call      89     2025-12-15

Pretty output (--pretty):

supertag transcript list --pretty
🎙️ Meetings with transcripts (3):

  ID            Meeting                         Lines  Date
  ───────────────────────────────────────────────────────────────
  M9rkJkwuED    Monthly Team Standup              127  2025-12-20
  xY7nKpQrSt    Q4 Planning Session               342  2025-12-18
  aB3cDeFgHi    Client Review Call                 89  2025-12-15

💡 Tip: Use 'supertag transcript show <id>' to view transcript

Options:

Flag Description
--limit <n> Maximum meetings to show (default: 20)
--pretty Formatted table output with header
--json Output as JSON
-w, --workspace <alias> Target specific workspace

Show Transcript Content

supertag transcript show <meeting-id>

Display the full transcript for a meeting:

Meeting: Monthly Team Standup
────────────────────────────────

[00:00:12] Speaker 1:
  Good morning everyone, let's get started with the standup.

[00:00:18] Speaker 2:
  Sure. Yesterday I finished the API integration and today
  I'm working on the frontend components.

[00:00:35] Speaker 1:
  Great progress. Any blockers?

Options:

Flag Description
--limit <n> Maximum lines to show (default: 100)
--pretty Formatted output with speaker sections
--json Output as JSON with full metadata
-w, --workspace <alias> Target specific workspace

JSON output includes:

  • Line ID
  • Speaker name (when available)
  • Start/end timestamps
  • Full text content

Search Within Transcripts

supertag transcript search "quarterly review"

Search for text that was spoken in meetings:

ID            Meeting                   Speaker      Text
───────────────────────────────────────────────────────────────────────────────
xY7nKpQrSt    Q4 Planning Session       Speaker 2    ...discussed the quarterly review...
M9rkJkwuED    Monthly Team Standup      Speaker 1    ...prep for quarterly review meeting...

Options:

Flag Description
--limit <n> Maximum results (default: 20)
--json Output as JSON
-w, --workspace <alias> Target specific workspace

Default Transcript Exclusion

Transcripts represent 6-10% of nodes in a typical Tana workspace but contain low-semantic-value spoken language (filler words, fragments, interruptions). By default:

  • supertag search excludes transcript lines
  • supertag search --semantic excludes transcript lines
  • supertag embed generate excludes transcripts

This keeps search results focused on your actual notes and documents.

Including Transcripts in Embeddings

If you want semantic search to include transcript content:

supertag embed generate --include-transcripts

This will embed all transcript lines, significantly increasing:

  • Embedding generation time (~90K additional nodes)
  • Storage requirements
  • Potential for "noisy" search results

MCP Tools

Three MCP tools are available for AI assistants:

Tool Description
tana_transcript_list List meetings with transcripts
tana_transcript_show Get transcript lines for a meeting
tana_transcript_search Search within transcript content

Example: Claude Desktop

"Find meetings where we discussed the pricing strategy"
→ Uses tana_transcript_search to find relevant discussions

"Show me the transcript from last week's planning meeting"
→ Uses tana_transcript_list to find the meeting
→ Uses tana_transcript_show to display the content

Use Cases

Find Where Something Was Discussed

# Search for a topic across all meeting transcripts
supertag transcript search "budget approval"

Review a Past Meeting

# List recent meetings
supertag transcript list --limit 10

# View specific meeting transcript
supertag transcript show M9rkJkwuED

Export Transcript for Analysis

# Export full transcript as JSON
supertag transcript show M9rkJkwuED --json > meeting-transcript.json

Check Meeting Coverage

# How many meetings have transcripts?
supertag transcript list --json | jq length

Technical Details

For information about how Tana stores transcripts internally, see Tana Transcript Structure.

Key points:

  • Transcripts are linked to meetings via the SYS_A199 field in metanodes
  • Each transcript line has optional speaker (SYS_A252) and timing (SYS_A253, SYS_A254) metadata
  • Timestamps use relative format (offset from meeting start)

Troubleshooting

Problem Solution
"No transcripts found" Ensure you have Tana meetings with transcription enabled
Empty speaker names Some transcripts only have generic "Speaker 1" labels
Missing timing info Older transcripts may lack timing metadata
Slow queries Large transcripts (500+ lines) may take 1-2 seconds