Quick Start¶
Your First Extraction¶
Point PikoClaw at any email archive:
PikoClaw auto-detects the format. Output goes to ./pikoclaw-output/wiki/.
What Gets Created¶
pikoclaw-output/
└── wiki/
├── index.md # Overview with stats table
├── contacts.md # Contact directory (sorted by activity)
├── contacts.json # Machine-readable graph with metrics
├── threads.md # Conversation threads by topic
├── calendar.md # Calendar events timeline
├── network-analysis.md # Graph intelligence (if networkx installed)
├── provenance.json # Cryptographic provenance metadata
└── emails/
├── all.md # Every email, newest first
├── inbox.md # Received messages
└── sent.md # Sent messages
Custom Output Directory¶
Multiple Sources¶
Merge multiple archives into one unified knowledge base:
PikoClaw threads messages across all sources and builds a single contact graph.
JSON Export¶
Generate machine-readable output for downstream tools, LLMs, or custom analysis:
This creates extraction.json alongside the wiki. For JSON only (no wiki):
Save Attachments¶
Extract binary attachment data:
Attachments are saved to wiki/attachments/msg-NNNNNN/ with sanitized filenames.
Limit Processing¶
For large archives, cap messages per source:
Example Output¶
After running pikoclaw extract mailbox.pst, open wiki/index.md:
# Knowledge Base: mailbox.pst
Extracted: 2026-02-23T15:30:00+00:00
PikoClaw v0.5.0
## Summary
| Metric | Count |
|--------|-------|
| Emails | 12,847 |
| Calendar Events | 234 |
| Contacts | 342 |
| Conversation Threads | 4,291 |
| Multi-message Threads | 1,847 |
The contacts, threads, and network analysis pages link from the index for easy navigation.