Pick an AI provider, paste your key, and make Quillin’ yours. Takes about five minutes the first time. After that, it just runs.
Quillin’ works with Claude, GPT, Gemini, Grok, Mistral, Groq, and local Ollama. Pick the one you already use, or start with Gemini or Groq: both have free tiers generous enough for real daily use.
Drafting a hard conversation, writing something that actually has to land: Claude follows nuance and your standing instructions better than anything else. Worth paying for when it counts.
GPT leans into absurdity in a way the others don't quite match. If you're going Pirate, Unhinged, or Victorian Offended, ChatGPT commits to the bit harder. Excellent for tone crimes.
Gemini's free tier is genuinely generous and it handles everyday rewrites well. Good default if you don't want to pay anything yet, or as a fast second opinion on something Claude already touched.
Google's Gemini models are fast and capable, and the free tier is genuinely generous for personal use. The easiest starting point if you want to try Quillin’ without spending anything.
Go to aistudio.google.com and sign in with your Google account.
Click the key icon in the bottom-left sidebar, or click Get API Key at the top of the page.
Click Create API Key, select a project (or create one), and copy the key.
Open Quillin’ Settings from the toolbar popup, paste it into the Gemini field, and save.
Open the extension popup → click the gear icon → paste into the Gemini field.
Groq isn't an AI model. It's a hardware company with blazing-fast inference. They run open-source models (like Meta's Llama) on custom chips. The result: very fast responses and one of the most generous free tiers around.
Go to console.groq.com and sign up.
Click API Keys in the sidebar, then Create API Key.
Copy the key. It starts with gsk_.
Open Quillin’ Settings from the toolbar popup, paste it into the Groq field, and save.
Open the extension popup → click the gear icon → paste into the Groq field.
Excellent at nuanced writing, tone adjustment, and following your standing instructions precisely. Haiku is fast and inexpensive; Sonnet is more capable for complex rewrites. New accounts get $5 in free credits, enough for thousands of checks.
Go to console.anthropic.com and sign in or create an account.
Click API Keys in the left sidebar.
Click Create Key, give it a name (e.g. "Quillin’"), and copy it. It starts with sk-ant-.
Open Quillin’ Settings from the toolbar popup, paste it into the Claude field, and save.
Open the extension popup → click the gear icon → paste into the Claude field.
Great for creative rewrites and tone transformation. GPT-4o mini is fast and cheap for everyday checks; GPT-4o is more powerful for complex rewrites. You'll need to add a payment method before your key will work.
Go to platform.openai.com and sign in.
Click your profile, then API Keys.
Click Create new secret key and copy it. You won't see it again, so don't close the tab.
Open Quillin’ Settings from the toolbar popup, paste it into the GPT field, and save.
Open the extension popup → click the gear icon → paste into the GPT field.
xAI's model: direct, fast, and a bit more opinionated than the others. Grok 3 mini is fast and affordable. Sign in with your X (Twitter) account.
Go to console.x.ai and sign in with your X account.
Click API Keys, then Create API Key.
Copy the key. It starts with xai-.
Open Quillin’ Settings from the toolbar popup, paste it into the Grok field, and save.
Open the extension popup → click the gear icon → paste into the Grok field.
A French AI company making strong, efficient models. Particularly good for multilingual writing and European languages. Mistral Small is one of the most cost-effective models available, great for high-volume use.
Go to console.mistral.ai and create an account.
Click API Keys in the sidebar.
Click Create new key and copy it.
Open Quillin’ Settings from the toolbar popup, paste it into the Mistral field, and save.
Open the extension popup → click the gear icon → paste into the Mistral field.
Ollama lets you run AI models locally. No internet, no API costs, complete privacy. Your text never leaves your machine. Requires a Mac, Linux, or Windows computer with enough RAM to run the model.
Download and install Ollama from ollama.com.
Open Terminal and run: ollama pull llama3.1 (or swap in any model you prefer).
Make sure Ollama is running. Look for the icon in your menu bar.
Open Quillin’ Settings and toggle Show Ollama models on. No key needed.
llama3.1, mistral, qwen2.5:7b. Larger models need more RAM but produce better results.Ollama must be running locally before Quillin’ can detect its models.
These are the tones that appear every time you run a check. You get up to five. Sixteen are built in. Write your own in plain words. Drag to reorder them in settings.
Describe the tone in plain words. "Email from a very disappointed but still polite British manager" works just as well as "Professional."
After a rewrite, all five of your styles appear at once. No waiting, no switching. Just pick the one that fits.
What you need on Monday isn't always what you need on Friday. Swap styles in settings and the next check reflects the new set.
When you're picking between five options quickly, a 🔥 is easier to spot than a word. Use them if that helps.
These are the rules that apply to every check, every time, on every provider. Think of it as your personal style guide baked in permanently. No repeating yourself, no reminding the AI.
Words and phrases here are never flagged as errors, regardless of what the AI thinks. One entry per line.
Quillin’ skips sites you add to the disabled list. The trigger button won't appear on them at all. You can also add sites by right-clicking anywhere on the page.
Key in, styles set, house rules written. Everything else is just using it.