🦉
"Who's who in my zoo?"
Most tools make you look things up.
WhosWhoZoo already knows.
It loads everything you've saved about people and answers from that — instantly and completely. Like talking to your own memory.
How It Works
💾Your memory, your files
Everything lives in 8 plain markdown files stored in Cloudflare KV — people, reflections, fragments, and loops (plus their 4 archive equivalents). You own the data. Export it anytime.
🧠Claude reads everything, every time
Every chat message loads all your active memory files into Claude's context. No vector search, no retrieval guessing — the full picture, every request. Ask anything.
📦Archive keeps it lean
Records you no longer need daily move to archive. Enable the Archive toggle when you need to search them. Archived files don't load in normal chat — keeping cost and speed optimal.
🎙Voice in, voice out
Tap the mic to enter conversation mode. Speak naturally. The app transcribes, sends to Claude, and reads the response aloud — with URLs and code replaced by verbal cues so it stays listenable.
Commands
Type / in the chat input to see autocomplete suggestions. All searches are fuzzy — you don't need exact names.
Search & Edit
/edit [query]
Open a record in the editor — browse and edit without spending tokens
free
/references [name]
Find every record that mentions a name or topic
free
/delete [query]
Permanently delete a record — requires typing "delete" to confirm
free
Journal
/journal <entry>
Polish and save a timestamped entry to reflections.md
tokens
/journal-flush [days]
Move entries older than N days to archive (default: 60)
free
Archive
/archive [query]
Move a record to long-term archive
free
/archive-restore [query]
Move an archived record back to active memory
free
/archive-review
AI suggests archive candidates · add [file] [deep] [criteria]
tokens
/archive-ignored
View or clear ignored archive suggestions
free
Compact
/compact [query]
Compact a specific record — AI removes redundancy
tokens
/compact-review
AI finds verbose sections · add [file] [deep] [deeper]
tokens
Memory Tools
/mnemonic [name]
Generate memory hooks for a person
tokens
/memory-game
People memory quiz — 5 questions, then continue?
tokens
Utility
/costs
Show today's and weekly Claude API cost estimate
free
/diagnostics
System health: memory sizes, API keys, cost log, errors
free
/r
Reset conversation context (memory stays loaded)
free
/cls
Clear chat history and context
free
/demo
Toggle demo mode — filters sensitive details from responses
free
/export
Download an encrypted backup (.whoszoo)
free
/export-plain
Download a plain-text zip (unencrypted)
free
/import
Restore memory files from an encrypted backup
free
/help
Show command list in chat
free
Header Controls
⚡
Haiku / Sonnet
Tap to switch AI model. Haiku is fast and cheap. Sonnet is smarter for complex tasks.
🎙
Voice mode
Full conversation mode — speak in, hear the response. Tap the orb to start/stop. Tap ✕ to exit.
📦
Archive
When ON, Claude includes archive files in its context, and /edit, /delete, /references all search archive too.
🎭
Demo
Filters sensitive details from responses. Useful when showing the app to someone.
⚙
Settings
Update your Worker URL or API key.
Tips
Just ask naturally. You don't need commands for most things. "Who is Adam?" or "what do I know about the PTMM project?" work as plain questions.
Single words trigger tag search. Type a single word or phrase with no verb and the app searches all records for that tag exhaustively before answering.
Add a file scope to any search. /edit people adam narrows to people.md. /archive loops archives from loops.md only.
Archive keeps Claude fast. Move stale records to archive. They're still searchable with the 📦 Archive toggle on — just not loaded in every chat message.
Voice skips URLs and code. In conversation mode, links say "see link in chat" and code says "see code in chat" — the full content is still in the chat bubble.
/r is your reset valve. If a conversation goes sideways, /r clears context without losing your memory files.