SerpApi
Give your local workflow optional live web results. Searches happen only when you request them and may consume SerpApi credits.
Drop the little demands in one quiet place. Gemma sorts what matters, what can wait, and what you can forget. All on your own computer.
Use Quiet Lab and the local Windows or Linux companion without an account. Sign in only if you want a task list saved to your Sloth Mode account across devices.
Optional cloud storage: tasks saved here are stored in your Sloth Mode account database. Private connector notes on Windows and Linux are never uploaded automatically.
Checking account status…
Connect your Windows or Linux companion to check RAM, CPU and graphics hardware. This check stays on your PC.
Downloads only after you explicitly approve Local AI setup. Your hardware profile is not sent to your Sloth Mode cloud account.
No installation, account or pairing required. With your permission, this browser downloads Gemma 4 E2B (2.01 GB) and runs it on your GPU using WebGPU. Your notes and prompts are not sent to Sloth Mode servers.
Checking your browser…
Preparing compatibility check…
The model stays in this browser's private storage until you remove it or the browser clears site data. The first download and the version-pinned AI runtime require internet access; offline reopening is not guaranteed yet. Browser inference does not monitor applications or replace native system controls.
Pair the Windows or Linux companion once, then work with your local notes here. No cloud model or account required.
Once paired, this area lets you manage your real local tasks and ask Gemma to sort them. Notes go directly from your browser to 127.0.0.1 on your own device, never through our servers.
Five minutes is enough to begin.
Check model readiness to begin.
The first model download needs your approval and uses several GB, depending on model quality. Once installed, the model runs on your PC through llama.cpp; no Ollama required.
Here's the idea, using three fictional notes. For real inference, choose Browser Gemma or pair the local companion. This walkthrough uses fixed examples.
Reply to my collaborator about the slides today
Finally learn how sourdough starters work
Keep the receipt from yesterday's train
Ready when you are.
Hit the button to see a sample sort.Send a short response while it's timely.
Set aside time whenever you're curious.
Keep for reference, no action needed.
Try your own fictional to-do list or a small chat excerpt. This live preview uses transparent offline rules, not a hosted model. Your input stays in this browser tab.
One task per line. Paste a short list, or import a .txt/.md export. No account or AI key needed here.
A focus window limits the number of suggested items, not their real completion time. Check urgency yourself. This preview never sends, deletes or archives anything.
Your source-linked suggestions will appear here. Nothing is decided for you.
The Windows connector stores notes locally and runs Gemma through llama.cpp, without Ollama. Notes and model requests stay on your computer.
Inspect the Gemma prompt and the code that checks each result against your notes. Adapt the open-source workflow to your needs.
Once the software and model have been downloaded, core sorting can run with no internet connection. Optional online helpers stay off until you choose them.
Install the Windows connector once. It runs in the notification area and the website becomes your interface.
Run SlothMode-Setup.exe from the Windows preview package. It installs the connector for your account, then opens Setup & Pairing. No Python or Ollama required.
Paste the setup key once on each trusted browser. Sloth Mode stores a device-bound signing key in that browser, not the original pairing secret. Chrome or Edge may ask for local network access.
Pair local app ↑Approve a model download of several GB, depending on model quality. Afterwards, inference runs locally through llama.cpp, even without a cloud AI account.
The installer preview is currently shared as a project download; a direct website download will follow Windows verification. The browser UI needs the site to load; data and Gemma inference remain on your computer. Quiet Lab always uses rules. Choose Browser Gemma or Plan with local Gemma for real inference.
Connect the tools you choose to use. Credentials are stored in your Windows companion, not in a Cloudflare database. Only connect what you need.
Give your local workflow optional live web results. Searches happen only when you request them and may consume SerpApi credits.
Link an ElevenLabs API key and check it from your PC. Narration and speech generation will be added in a later release.
Research and memory-assisted workflows for open models. No live connector integration yet.
Explicitly approved website automation through Stagehand. Not connected in this version.
Local error diagnostics and agent traces are planned. Sentry is not connected in this version.
An optional second-opinion model for research and documents. Not connected in this version.
Your query is sent directly from the Windows companion to SerpApi. Nothing is searched automatically.
Your brain is for living.
Not remembering everything.