# Check if qmd is already installed
which qmd

# Check if Bun is available (required)
which bun

# macOS only: Check SQLite via Homebrew
brew list sqlite 2>/dev/null && echo "SQLite OK" || echo "Need: brew install sqlite"

# Check available disk space (need ~2.5GB)
df -h ~ | tail -1

═══════════════════════════════════════════════════════════════════════
                    SEMANTIC SEARCH FOR YOUR VAULT
═══════════════════════════════════════════════════════════════════════

What is this?
─────────────
Right now, finding notes requires knowing the exact words used.
Search "product-led growth" — won't find notes saying "PLG" or
"self-serve motion".

Semantic search understands meaning, not just keywords. It finds
conceptually related content even when terminology differs.


How does it work?
─────────────────
Your notes get converted to "embeddings" — mathematical
representations of meaning. When you search, your query becomes
an embedding too. The system finds notes whose meaning is close
to your query's meaning.

Think of it like: instead of matching letters, we're matching ideas.


What gets installed?
────────────────────
Three small AI models run locally on your machine:

  MODEL                      PURPOSE                    SIZE
  ─────────────────────────────────────────────────────────────
  EmbeddingGemma-300M        Converts text to meaning   ~300MB
                             vectors. The core
                             "understanding" model.

  Qwen3-Reranker-0.6b       Re-orders results by       ~640MB
                             true relevance. Improves
                             result quality.

  QMD-Query-Expansion-1.7B   Expands your search to     ~1.1GB
                             include related terms.
                             "PLG" → also searches
                             "product-led", "freemium"

  Total: ~2GB one-time download


Privacy & Security
──────────────────
- Everything runs locally — your notes never leave your machine
- No API keys required
- No cloud services
- Models downloaded from HuggingFace (open source)
- Index stored in ~/.cache/qmd/ (not in your vault)


What changes in your workflow?
──────────────────────────────
Enabling semantic search silently upgrades these skills:

  /daily-plan    — Enriches meeting prep with thematically
                   related past discussions

  /meeting-prep  — Discovers past discussions related by
                   meaning, not just name matching

  /triage        — Matches inbox items to goals by meaning,
                   catches semantic duplicates

  Person Lookup  — Finds "the VP of Sales mentioned..." even
                   without a name

  Search & Recall — All vault searches use hybrid retrieval
                    (BM25 + vectors + LLM reranking)


System Requirements
───────────────────
  ~2.5GB disk space (for models + index)
  macOS: Homebrew SQLite required (brew install sqlite)
  Bun runtime (will install if missing)

# Install Bun if missing
if ! command -v bun &> /dev/null; then
    echo "Installing Bun runtime..."
    curl -fsSL https://bun.sh/install | bash
fi

# macOS: Install SQLite if missing
if [[ "$OSTYPE" == "darwin"* ]]; then
    if ! brew list sqlite &> /dev/null; then
        echo "Installing SQLite via Homebrew..."
        brew install sqlite
    fi
fi

# Install qmd globally
echo "Installing qmd..."
bun install -g github:tobi/qmd

echo "Downloading AI models (~2GB)..."
echo "This happens once. Future searches are instant."
echo ""
echo "Models downloading:"
echo "  EmbeddingGemma-300M  — Converts text to meaning vectors"
echo "  Qwen3-Reranker-0.6b  — Improves result relevance"
echo "  QMD-Query-Expansion   — Expands searches with related terms"
echo ""

# Trigger model download with a simple embed operation
cd "$VAULT_PATH" && qmd embed --help 2>/dev/null || true

node "$VAULT_PATH/.scripts/semantic-search/scan-vault.cjs"

{
  "candidates": [
    {
      "name": "people",
      "path": "05-Areas/People",
      "glob": "**/*.md",
      "fileCount": 23,
      "context": "Person pages with meeting history, relationship notes, action items, and role context",
      "benefit": "Person lookup finds references by role/title, not just name",
      "example": "Search 'VP of Sales' finds the person even if their name isn't mentioned"
    },
    {
      "name": "accounts",
      "path": "05-Areas/Companies",
      "glob": "**/*.md",
      "fileCount": 8,
      "context": "Company and account pages with deal status, relationship notes, and interaction history",
      "benefit": "Deal prep pulls account-specific context without inbox noise",
      "example": "Search 'renewal risk' finds accounts with churn signals"
    }
  ],
  "skipped": [
    {
      "name": "career",
      "reason": "Folder not set up (run /career-setup)",
      "path": "05-Areas/Career"
    }
  ],
  "totalFiles": 156,
  "totalCandidates": 6
}

I scanned your vault and found candidates for [N] smart collections:

  COLLECTION     FILES   WHAT IT ENABLES
  ─────────────────────────────────────────────────────────────────
  people         23      Person lookup finds references by
                         role/title, not just name. Search
                         "VP of Sales" finds the person even
                         if their name isn't mentioned.

  meetings       47      "What was discussed about X?" searches
                         meeting notes specifically, not your
                         entire vault.

  tasks          1       Smart task matching for triage. Catches
                         semantic duplicates like "Review metrics"
                         ≈ "Check quarterly numbers".

  projects       8       Project health pulls related docs
                         without inbox noise.

  goals          1       Goal alignment finds thematically
                         related work across your vault.

  priorities     1       Weekly planning discovers patterns
                         in past priority choices.

  Skipped (not enough content yet):
  - accounts  — No company pages found (create some first)
  - content   — No content files yet
  - career    — Career folder not set up (run /career-setup)

  Create all [N] collections? [Y]
  Pick specific ones?         [P]
  Just create a basic index?  [B]

# Create the collection
qmd collection add "$VAULT_PATH/<path>" --name <name> --mask "<glob>"

# Add semantic context (helps the reranker understand what's in the collection)
qmd context add <name> "<context description>"

echo "Creating embeddings for all collections..."
qmd embed

{
  "mcpServers": {
    "qmd": {
      "type": "stdio",
      "command": "qmd",
      "args": ["mcp"]
    }
  }
}

{
  "mcpServers": {
    "qmd": {
      "command": "qmd",
      "args": ["mcp"]
    }
  }
}

═══════════════════════════════════════════════════════════════════════
                    SEMANTIC SEARCH ENABLED
═══════════════════════════════════════════════════════════════════════

Your vault now has smart, meaning-aware search.

Collections created:
  [list each collection with file count]

Try it:
  qmd query "product led growth strategies"
  qmd query "customer churn patterns" -c accounts
  qmd search "meeting with Sarah" -c meetings

What's different now:
  /daily-plan    — meeting context enriched with thematic connections
  /meeting-prep  — attendee lookup finds role/title references
  /triage        — semantic routing with goal-aware matching
  Person Lookup  — finds "the VP of Sales" without needing the name
  Search & Recall — hybrid retrieval (BM25 + vectors + reranking)

Index updates automatically when you run: qmd update
To check status anytime:                  qmd status
To force full rebuild:                     qmd embed -f

As your vault grows, I'll suggest new collections when there's
enough content to benefit. Run this skill again anytime to check.

node "$VAULT_PATH/.scripts/semantic-search/scan-vault.cjs" --health-check

{
  "existing": ["people", "meetings", "tasks"],
  "newCandidates": [
    {
      "name": "accounts",
      "path": "05-Areas/Companies",
      "fileCount": 5,
      "reason": "You've created 5 company pages since last check"
    }
  ],
  "staleCollections": [
    {
      "name": "meetings",
      "lastUpdated": "12 days ago",
      "currentFiles": 52,
      "indexedFiles": 47,
      "drift": 5
    }
  ],
  "pendingEmbeddings": 8,
  "suggestions": [
    "Run 'qmd update' to re-index changed files",
    "Run 'qmd embed' to embed 8 pending documents"
  ]
}

SEMANTIC SEARCH HEALTH CHECK
─────────────────────────────────────────────────────────────────

Active Collections:
  people      54 files    Updated 2h ago     Healthy
  meetings    47 files    Updated 12d ago    ⚠ Stale (5 new files)
  tasks        4 files    Updated 1d ago     Healthy

New Collection Candidates:
  accounts    5 files     You've built enough company pages for
                          a dedicated collection. This means deal
                          prep will pull account context specifically.

                          → Create accounts collection? [Y/n]

Maintenance:
  8 documents need embedding (run 'qmd embed')
  meetings collection is 12 days stale (run 'qmd update')

Quick fix: qmd update && qmd embed

node "$VAULT_PATH/.scripts/semantic-search/scan-vault.cjs" --suggestions-only

💡 Semantic search suggestion: You now have 5 company pages.
   Want me to create an 'accounts' collection? This means deal
   prep will search account context specifically.
   Run /enable-semantic-search to set it up.

// In your skill or hook
const { execSync } = require('child_process');

function isSemanticSearchAvailable() {
  try {
    execSync('which qmd', { stdio: 'pipe' });
    const status = execSync('qmd status', { stdio: 'pipe' }).toString();
    return status.includes('Documents');
  } catch {
    return false;
  }
}

// Use in search logic
if (isSemanticSearchAvailable()) {
  // Use qmd query for semantic search
  const results = execSync(`qmd query "${query}" -c people`, { stdio: 'pipe' });
} else {
  // Fall back to grep/file search
}

const { checkSemanticSearch } = require('.scripts/semantic-search/check-availability.cjs');
const status = checkSemanticSearch();
if (status.available) {
  // Use semantic search
}

# Remove index
qmd cleanup

# Remove the qmd binary
bun remove -g qmd

# Remove MCP config (manual — delete the "qmd" entry from your config)