Slate

Like many of you, I’ve been dangerously obsessed with LLMs ever since I first used ChatGPT. I’m pretty sure I've stress-tested every token and request limit on ChatGPT, Claude, and a dozen others.

When GPT-4 dropped, I quit my perfectly stable job as a B2B SaaS product manager and dove headfirst into programming — with the constant reassurance from ChatGPT that I was doing “great, buddy! You’ve made only 966 errors this time!”

For months, my workflow looked like digital chaos:

• ChatGPT in one tab

• Claude in another (for depth, when I could get past the rate limits)

• Google AI Studio attempting to digest giant codebases

• Notion, Docs, Obsidian, and 6+ other apps to organize my spaghetti-brain

Token limits would always kill my flow just as things were getting good. The LLMs were powerful, but managing them felt like asking a bunch of squirrels to maintain a filing cabinet. Claude Projects helped, but still felt meh UX-wise.

I tried Copilot and Cursor too, but as a new coder, it felt like teaching ballroom dancing to an octopus. (Apologies to any octopus dancers reading this.) I wanted control, not autocomplete. And I wanted my context to stay alive.

So I built Slate.

It’s kind of like if Notion and ChatGPT had a baby — and that baby had a photographic memory and access to multiple models.

Slate lets you:

• Use up to 1 million tokens per session (~750k words)

• Add side-by-side notes, files, and tasks while you chat

• Talk to GPT-4, Claude, Gemini, and others from one interface

• Track token usage and model costs directly

• Avoid losing your mind switching tabs every 10 minutes

Feature Comparison (a.k.a. Why I fired ChatGPT for my own tool):

1M+ token context

• ❌ ChatGPT/Claude: Nope

• ✅ Slate: Yep, seriously

Persistent memory

• ❌ Goldfish vibes

• ✅ Elephant brain

Multiple model access

• ❌ Single-vendor loyalty

• ✅ AI buffet, your pick

Knowledge management

• ❌ Uh, where was that convo again?

• ✅ All your notes, files, chats — together

Pay per token

• ❌ Hidden pricing, weird limits

• ✅ Full control and usage visibility

Model Options

Paid Models:

• Use OpenAI, Anthropic, Google, etc.

• Your data stays private

• Pay-per-token, high reliability

Experimental Models:

• Free, newer open models

• Your data may be retained

• Good for exploration, prototyping, or just vibing

🧠 Ideal Use Cases (in handy list form — because organization!)

Programming:

• Paid: Private codebases, company secrets, things that should not spontaneously refactor themselves.

• Experimental: Open-source tinkering, exploratory projects, side quests that might spawn a self-aware debugger.

Product Management:

• Paid: Confidential roadmaps, secret product plans, your boss’s OKR to leverage blockchain for Slack channel naming conventions.

• Experimental: Early-stage brainstorming, throwing spaghetti at walls, building features no one asked for.

Content Creation:

• Paid: Proprietary articles, confidential drafts, and that “just-for-fun” side project called world_domination_final_FINAL.docx.

• Experimental: Public blogs, exploratory content, your unpublished Harry Potter fanfic.

Learning & Research:

• Paid: Sensitive studies, private research, embarrassing questions you don’t want Google knowing about.

• Experimental: General learning, exploring the latest AI models, asking models to explain string theory using sandwiches.

🎁 The Giveaway (a.k.a. My API credit panic)

I’ve got $300 worth of Gemini credits that expire in 48 hours. So I’m giving away 1 million in-app tokens to every new user who signs up right now.

No strings. Just try it, explore the Gemini models (please), and let me know what breaks. I’d rather see it used by curious humans than disappear into the void.

👉 Try it here: www.slate-app.io

Thanks for reading. This is my first “coding” project, so be gentle 🙏