Wallie watches your screen, hears your audio, and reacts in real-time — with a personality you design in plain English. Download it, give it a character, watch it come alive.
Wallie doesn't narrate — it experiences. Sight and sound fuse into one coherent reaction, the way a real person takes in a moment.
Watches your screen and reacts in first person. Knows if you're gaming, browsing, or watching — and reacts like it's there, never like a screen reader.
Hears music, videos and voices on your system. Reads a song's mood, tempo and lyrics — calls a track a banger or roasts a muddy mix.
NewDescribe a personality in plain English — rage gamer, film snob, conspiracy theorist — and Wallie becomes it. No code, no limits.
Someone told it to play Minecraft as an overconfident conspiracy theorist — so it called the villagers government agents and the Ender Dragon a secret organization. You describe a personality, it commits.
They describe your screen like a help desk. They repeat themselves constantly. They ask "what do you think, chat?" into the void. They're tech demos, not shows.
"I can see a YouTube homepage with several video thumbnails arranged in a grid layout."
"okay let me pull this up, I saw someone post about this earlier and I have thoughts."
Says the same observation with slightly different words every 30 seconds.
Bigram/trigram similarity scoring, phrase cooldowns, and rolling history catch every repeat.
Awkward 3-second gaps between every thought. Choppy, robotic pacing.
Mood-driven tempo, pipeline overlap, silence beats, and thought development across minutes.
One framework. Infinite characters. Every component is swappable.
Name, energy, humor style, catchphrases, running gags, banned words, opinions, backstory. Describe a character in plain English and it becomes it.
Screen capture with activity classification. Knows when you're gaming, browsing, or watching — and reacts in first person, not narration.
Hears system audio and reads music — mood, tempo, key, lyrics — fused with vision into one reaction. Sight and sound, one voice.
Rolling summarizer compresses old context. Cross-session memory persists facts. No degradation over time.
Spectral lip sync shapes the mouth per sound, plus blink, body motion, and mood-reactive expressions over VTube Studio.
Configure everything from a web UI. No YAML, no terminal after setup. Live preview and instant testing.
Identity, personality, voice, topics, vision, hearing, chat, avatar — every parameter lives in one dashboard. Save profiles, switch characters, go live.
Identity — who your streamer is
Personality — how they talk, not what they talk about
No Git, no terminal. Download the folder, run one file — it installs itself and opens in your browser.
Grab the Wallie folder — one click, no GitHub account needed.
Unzip, then double-click start.bat (Windows) or run start.sh (Mac/Linux). It auto-installs everything.
The dashboard opens in your browser. Pick a personality, add a key, hit Start.
Windows · macOS · Linux · Python is installed automatically if you don't have it
Wallie runs on your machine with your own API keys. No subscription, no middleman, no markup.
New features, one-click installers, and the next personality drops — no spam, just the good stuff.
We'll only email about Wallie. Unsubscribe anytime.
Download it. Give it a personality. Watch it react to everything.