Initial commit — memex

A compounding LLM-maintained knowledge wiki.

Synthesis of Andrej Karpathy's persistent-wiki gist and milla-jovovich's
mempalace, with an automation layer on top for conversation mining, URL
harvesting, human-in-the-loop staging, staleness decay, and hygiene.

Includes:
- 11 pipeline scripts (extract, summarize, index, harvest, stage,
  hygiene, maintain, sync, + shared library)
- Full docs: README, SETUP, ARCHITECTURE, DESIGN-RATIONALE, CUSTOMIZE
- Example CLAUDE.md files (wiki schema + global instructions) tuned for
  the three-collection qmd setup
- 171-test pytest suite (cross-platform, runs in ~1.3s)
- MIT licensed

.gitignore

@@ -0,0 +1,35 @@
+# Conversation extraction state — per-machine byte offsets, not portable
+.mine-state.json
+
+# Log files from the mining and maintenance pipelines
+scripts/.mine.log
+scripts/.maintain.log
+scripts/.sync.log
+scripts/.summarize-claude.log
+scripts/.summarize-claude-retry.log
+
+# Python bytecode and cache
+__pycache__/
+*.py[cod]
+*$py.class
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+
+# Editor / OS noise
+.DS_Store
+.vscode/
+.idea/
+*.swp
+*~
+
+# Obsidian workspace state (keep the `.obsidian/` config if you use it,
+# ignore only the ephemeral bits)
+.obsidian/workspace.json
+.obsidian/workspace-mobile.json
+.obsidian/hotkeys.json
+
+# NOTE: the following state files are NOT gitignored — they must sync
+# across machines so both installs agree on what's been processed:
+#   .harvest-state.json   (URL dedup)
+#   .hygiene-state.json   (content hashes, deferred issues)