What problem does it solve?
Import Hyprnote meeting transcripts from the Hyprnote app into the appropriate project context repositories, enabling centralized accessibility and historical continuity.
Core Features & Use Cases
- Automatically transfer Hyprnote transcripts into per-project context repositories.
- Classify sessions by project keywords and route them to the correct context folder.
- Keep context repos up to date by periodically importing new transcripts and preserving metadata.
Quick Start
Place Hyprnote transcripts in the Hyprnote app and run the import-hyprnote skill to populate per-project context repositories.
When to Use
- After recording meetings in Hyprnote
- When transcripts need to be sorted to project-specific data directories
- Periodically to keep context repos up to date
Configuration
Requires .entourage/config.json with project definitions:
{
"projects": [
{
"name": "entourage-context",
"path": "~/Documents/code/@orgs/my-entourage/entourage-context",
"description": "AI agent development and product specs",
"keywords": ["agent", "spec", "inbox", "timeline", "Jared"]
}
]
}
Workflow
Step 1: Load Configuration
- Read
.entourage/config.json from current working directory
- If not found, prompt user to create configuration:
- Scan
~/Documents/ for *-context/ directories
- Present found projects for confirmation
- Ask for description and keywords for each
- Save to
.entourage/config.json
Step 2: Discover Hyprnote Sessions
- List all sessions in
~/Library/Application Support/hyprnote/sessions/
- For each session, check for:
_meta.json (required)
_transcript.json (required)
_memo.md (optional)
- Skip sessions that are empty (no transcript content)
Step 3: Classify Sessions
For each session:
- Read
_transcript.json and reconstruct text:
words.map(w => w.text).join('').trim()
- Read
_memo.md if exists
- Match against project keywords (case-insensitive)
- Assign to best matching project, or "unclassified" if no match
Step 4: Present Classification for Approval
Present a table to the user:
## Session Classification
| Session | Date | Title | Assigned To | Confidence |
|---------|------|-------|-------------|------------|
| abc123 | 2026-01-14 | Team standup | entourage-context | High (5 keywords) |
| def456 | 2026-01-13 | Procurement call | viran-context | Medium (2 keywords) |
| ghi789 | 2026-01-12 | Personal notes | unclassified | Low (0 keywords) |
Proceed with this classification? (You can reassign any session)
Wait for user confirmation or adjustments.
Step 5: Execute Import
For each confirmed session:
- Create destination directory:
{project_path}/data/transcripts/hyprnote/{session_id}/
- Copy files:
_meta.json
_transcript.json
_memo.md (if exists)
- Do NOT delete source files (user can clean up manually)
Step 6: Commit Changes (Optional)
For each affected project:
- Ask user if they want to commit the changes
- If yes:
git -C {project_path} add data/transcripts/hyprnote/
git -C {project_path} commit -m "Import Hyprnote transcripts from {date}"
Output
Summary of what was imported:
## Import Complete
Imported 3 sessions:
- entourage-context: 2 sessions
- viran-context: 1 session
Skipped:
- 1 unclassified session (ghi789)
- 2 empty sessions
Run `/update-timeline` in each context repo to process these transcripts.
Error Handling
- No config found: Guide user through setup wizard
- No Hyprnote sessions: Report "No sessions found in Hyprnote"
- Project path doesn't exist: Warn and skip that project
- Git not available: Skip commit step, warn user
Execution Behavior
This skill returns results to the calling context and does NOT stop execution.