Jottings•me

Dashboard →

How AI Knows Your Jot Terminology

• 5 min read • AI Features

How AI Knows Your Jot Terminology

When you ask an AI assistant to help you with your Jottings site—whether it's editing a jot, brainstorming content, or improving your site description—you want that assistant to understand your world without you explaining everything from scratch.

But here's the problem: AI models are trained on general internet data. They know what blogs are. They know what microblogs are. But they don't inherently know what your jots are, what makes your creator voice unique, or how you use Jottings to publish.

So we built something simple but powerful into Jottings: contextual awareness through prompt injection.

What Is Prompt Injection (The Good Kind)?

When I say "prompt injection," I'm talking about the intentional kind—not the security vulnerability you read about in the news. We're injecting helpful context into prompts you send to AI assistants so they understand your Jottings ecosystem without you having to explain it every time.

Think of it like giving an AI colleague a quick briefing before a meeting. Instead of them guessing, they know:

  • What a "jot" is in your world
  • Your site's purpose and focus
  • Your author bio and voice
  • The tags you use
  • How your site is organized

Where Does This Context Come From?

Your Jottings dashboard already contains rich information about your site:

Site Settings — your author bio, site description, what you write about Jot History — themes, tone, topics, and patterns in your writing Tags — the categories and organizational structure you've created Site Configuration — how your site appears to visitors

When you copy a jot to share with Claude, ChatGPT, or another AI assistant, Jottings can bundle this metadata alongside your text. We're not sending your entire jot history to AI companies (that would be a privacy nightmare). Instead, we're sending a curated summary: your creator voice, your site's focus, and relevant terminology.

How It Actually Works

Here's a simplified example. Without context injection, when you paste a jot into ChatGPT and ask for help, the assistant sees:

"I built a platform for creators. The alpha launch went well but
I'm worried about scaling."

The assistant might give generic startup advice. Fine, but not tailored to you.

With context injection, the assistant sees:

[CREATOR CONTEXT]
Site: Jottings
Author: Vishal
Bio: Founder of Jottings, writing about building
indie products and serverless architecture
Tags: #product-development #aws #entrepreneurship
Voice: Direct, personal, technical but accessible
Platform Terminology:
- Jot = Short-form piece of content on Jottings
- Site = Personal microblog subdomain (e.g., vishal.jottings.me)
- Build = Static site generation process
- Subscriber = Reader of your Jottings site
[END CREATOR CONTEXT]

I built a platform for creators. The alpha launch went well but
I'm worried about scaling.

Now Claude understands you're writing about an indie product, that you have a technical audience, and that you care about reliability and scalability. The advice shifts from generic to contextual.

Why This Matters for Creators

Most AI assistants are optimized for professional writing or general audiences. But creators on Jottings have unique needs:

Maintaining your voice — Your jots have a distinct tone. Injecting your bio and writing history helps AI preserve that voice when you ask for editing help.

Domain expertise — If you write about technical topics, niche hobbies, or specialized knowledge, context injection ensures the AI doesn't dumb down the content.

Platform literacy — When you ask "How should I structure this as a jot?", the AI understands that jots are concise, personal, and meant for your dedicated audience—not 500-word blog posts.

Consistent terminology — You use specific words, frames, and concepts. The AI learns them and uses them back, reducing friction in your workflow.

The Privacy-First Approach

Here's what we don't do:

  • We don't send your full jot history to OpenAI, Anthropic, or anyone else
  • We don't permanently store AI conversation history
  • We don't log what you ask AI assistants to do
  • We don't train models on your content

The context injection is entirely in your control. If you're sharing a jot with an AI, you decide whether to include your creator context. It's similar to how you might open a Google Doc with comments—you're actively choosing what information to share.

Looking Forward

Right now, context injection is mainly useful when you're manually copying jots to share with AI assistants. But I see opportunities to go deeper:

  • AI-assisted jot writing — We could build a tool directly in Jottings that understands your voice and helps you draft
  • Smart tag suggestions — AI could recommend tags based on your historical patterns
  • Content optimization — AI could analyze your jots and suggest improvements while preserving your voice
  • Subscriber insights — Help you understand which themes resonate most with your audience

Each of these would rely on the same principle: keeping your context front-and-center so AI gets you, not just your words.

The Bigger Picture

As creators, we're outsourcing more cognitive work to AI. Writing, editing, brainstorming, formatting—these are increasingly AI-assisted tasks. But AI is only useful if it understands what makes you you.

Jottings is betting that the future of creator tools isn't about AI doing everything for you. It's about AI understanding your world well enough to be a thoughtful collaborator.

A collaborator who knows what a jot is. Who knows your voice. Who understands why you publish the way you do.

If you're publishing on Jottings and want your AI assistants to understand your creator context, start by filling out your site settings with a detailed bio, description, and tags. The richer your creator profile, the better the context your AI collaborators will have.

That's the whole idea: your tools should understand your world.

← Back to Blog