Recall Local

Indexes your entire ~/clawd/memory/ archive and serves it as a searchable web UI + API at http://localhost:3456. No external services, no API keys, nothing leaves your machine.

Setup (first time)

# Copy the server to your tools directory
mkdir -p ~/clawd/tools/recall-local
cp "$(dirname "$0")/scripts/server.js" ~/clawd/tools/recall-local/server.js

# Create a LaunchAgent so it starts on login
cat > ~/Library/LaunchAgents/ai.wren.recall-local.plist << 'EOF'
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
  <key>Label</key><string>ai.wren.recall-local</string>
  <key>ProgramArguments</key>
  <array>
    <string>/opt/homebrew/bin/node</string>
    <string>/Users/YOUR_USERNAME/clawd/tools/recall-local/server.js</string>
  </array>
  <key>RunAtLoad</key><true/>
  <key>KeepAlive</key><true/>
  <key>StandardOutPath</key><string>/Users/YOUR_USERNAME/clawd/tools/recall-local/recall.log</string>
  <key>StandardErrorPath</key><string>/Users/YOUR_USERNAME/clawd/tools/recall-local/recall.log</string>
</dict>
</plist>
EOF

# Replace YOUR_USERNAME, then load it
launchctl load ~/Library/LaunchAgents/ai.wren.recall-local.plist

Or just run it manually: node ~/clawd/tools/recall-local/server.js &

Search (agent use)

curl -s "http://localhost:3456/search?q=YOUR+QUERY" | python3 -c "
import json,sys
d = json.load(sys.stdin)
print(f'{d[\"total\"]} chunks indexed')
for r in d['results'][:5]:
    print(f'[{r[\"source\"]}] {r[\"text\"][:300]}')
    print()
"

Check if running / start if down

curl -s http://localhost:3456/search?q=test > /dev/null 2>&1 && echo "UP" || (echo "Starting..."; launchctl start ai.wren.recall-local; sleep 2)

Human UI

Open http://localhost:3456 in a browser. Search box, results below. Works on mobile too if on same local network.

Tips

• Reloads all files on every search — always reflects the latest memory
• Source filenames are dates (2026-03-01.md) — useful for spotting when something happened
• Multi-word queries score better than single words
• For broad questions use general terms; for specific lookups use exact terms