/quick-capture

Quick Capture - SQLite Notes Ingestion

Ingest loose notes into the SQLite notes database at .sqlite/notes.db.

Arguments

• $ARGUMENTS - The raw notes to capture (can be messy, hierarchical, bullet points, stream of consciousness)

Process

1. Analyze and Chunk

Read the input and identify distinct ideas, not fragments. The goal is index cards (zettelkasten), not post-it notes.

Index card test: Can this note stand alone and say something meaningful? Does it have enough context to be useful when encountered later?

Too granular (post-it):

"Less state machine, more differential equations"

Right granularity (index card):

"Less state machine, more differential equations - the system evolves continuously based on conditions, not discrete transitions. Small changes to ontology yield large effects because you're shaping a possibility space, not defining fixed states."

2. Propose Chunking to User

Before creating notes, present the proposed breakdown:

• List each proposed note with a title and 1-sentence summary
• Identify theme clusters
• Ask user to confirm or adjust granularity

3. Create Notes

For each approved chunk, create a note in the inbox:

sqlite3 .sqlite/notes.db <<'SQL'
INSERT INTO notes (id, title, body, folder, origin, tags, captured_at)
VALUES (
  'N-' || strftime('%Y%m%d', 'now') || '-' || lower(hex(randomblob(4))),
  'Title here',
  'Content here',
  'inbox',
  'me',
  json_array('tag1', 'tag2'),
  strftime('%Y-%m-%dT%H:%M:%SZ', 'now')
);
SQL

Note ID format: N-YYYYMMDD-xxxxxxxx (date + 8 random hex chars)

4. Link Related Notes

Query existing notes to find connections:

# List inbox notes
sqlite3 -header -column .sqlite/notes.db "SELECT * FROM v_inbox;"

# All notes with id and title
sqlite3 -header -column .sqlite/notes.db "SELECT id, title FROM notes ORDER BY captured_at DESC;"

Create links between related notes:

sqlite3 .sqlite/notes.db <<'SQL'
INSERT INTO links (source_id, target_id, rel_type)
VALUES ('N-20260208-a1b2c3d4', 'N-20260207-e5f6g7h8', 'linksTo');
SQL

5. Offer Breadcrumb Generation

After capture, ask the user if they want a breadcrumb generated. If yes:

1. Query all recent notes
2. Analyze themes, connections, questions, and momentum
3. Create a breadcrumb:

sqlite3 .sqlite/notes.db <<'SQL'
INSERT INTO breadcrumbs (
  id, summary, themes, connections, questions, momentum,
  window_start, window_end, notes_considered, prev_breadcrumb_id, generated_at
)
VALUES (
  'BC-' || strftime('%Y%m%d', 'now') || '-' || lower(hex(randomblob(4))),
  'Summary of recent thinking...',
  json_array('theme1', 'theme2'),
  'How ideas connect...',
  json_array('Question 1?', 'Question 2?'),
  'exploring',
  strftime('%Y-%m-%dT%H:%M:%SZ', date('now', '-7 days')),
  strftime('%Y-%m-%dT%H:%M:%SZ', 'now'),
  5,
  (SELECT id FROM breadcrumbs ORDER BY generated_at DESC LIMIT 1),
  strftime('%Y-%m-%dT%H:%M:%SZ', 'now')
);
SQL

4. Batch link breadcrumb to analyzed notes:

sqlite3 .sqlite/notes.db <<'SQL'
INSERT INTO links (source_id, target_id, rel_type)
SELECT
  (SELECT id FROM breadcrumbs ORDER BY generated_at DESC LIMIT 1),
  notes.id,
  'analyzedNotes'
FROM notes
WHERE folder = 'inbox'
  AND captured_at >= date('now', '-7 days');
SQL

Key Principles

• Capture-heavy, not pristine: Get ideas in, refine later
• Origin tracking: Always set origin: me for user content, origin: llm for AI-generated
• Inbox first: Everything starts in inbox, moves to working/permanent during review
• Links over tags: Explicit links between notes are more valuable than shared tags
• Breadcrumbs for continuity: Regular breadcrumbs maintain thinking trails across sessions

Useful Commands

# List inbox (use view)
sqlite3 -header -column .sqlite/notes.db "SELECT * FROM v_inbox;"

# Delete a note
sqlite3 .sqlite/notes.db "DELETE FROM notes WHERE id = 'N-20260208-a1b2c3d4';"

# Find notes by tag
sqlite3 -header -column .sqlite/notes.db <<'SQL'
SELECT id, title, tags FROM notes
WHERE EXISTS (SELECT 1 FROM json_each(tags) WHERE value = 'theme');
SQL

# Full-text search
sqlite3 -header -column .sqlite/notes.db <<'SQL'
SELECT n.id, n.title, snippet(notes_fts, 1, '**', '**', '...', 30) AS snippet
FROM notes_fts
JOIN notes n ON notes_fts.rowid = n.rowid
WHERE notes_fts MATCH 'search term'
ORDER BY rank;
SQL