Give Your AI Eyes

Yidam provides real-time app context to your LLM — so users get helpful, contextual responses without complex prompt engineering.

Works with any LLM • Use your existing API keys

Why should only big teams get
smart AI UX?

Notion, GitHub Copilot, and Cursor all built deeply integrated AI that understands what users are doing.

Now you can too — without the custom infrastructure.

Yidam gives your AI that same context-rich awareness —
in just a few lines of code.

Notionfor docsCursorfor codeYidamfor everything else

The Problem Every Developer Knows

Your AI is blind. It doesn't know:

  • What screen the user is on
  • What they've clicked or selected
  • What data they're looking at
  • What they're trying to accomplish

The Result?

Chat widgets that ask "What page are you on?"

AI copilots that need manual context

Smart features that aren't very smart

Form assistants that can't see existing values

The Real Cost

Hours of prompt engineering
Frustrated users
Generic AI responses
Missed automation opportunities

How Yidam Works

1. Sync Your App State (3 lines)

const [currentScreen] = yidam.sync('screen', 'checkout');const [selectedData] = yidam.sync('selectedRow', userData);const [formValues] = yidam.sync('form', currentForm);

2. Make Any AI Call

const response = await yidam.ask("Help with my order");const action = await yidam.complete("Generate report");const result = await yidam.enhance(yourLLMCall);

3. AI Automatically Knows Context

Current screen/page
Selected data
Form values
User actions
App state

Result: AI responses that actually make sense

What Your AI Is Trying (and Failing) To Do

Here's what LLMs are being asked to do — often without real-time context

1. Answer questions about the app

How devs solve this now:

  • Embed docs into vector DB
  • Run RAG on every query
  • Hope it finds the right section

What's missing:

They don't know where the user is or why they're asking.

"You're on Team Settings — here's docs for managing roles"

2. Summarize or explain current data

Today:

LLMs get a blob of data and try to make sense of it

With Yidam:

"They're looking at user ID 342 with 3 declined payments"

Now the summary can be precise

3. Fill out or edit something

Today:

LLMs don't see pre-filled values unless manually included

With Yidam:

"The form has 'John Doe' — offer autocomplete, don't overwrite"

4. Take actions / generate commands

Today:

LLMs hallucinate actions, devs must validate against state

With Yidam:

"They're in Reporting with Q4 selected — just run export"

A Message from Your AI:

"If I'm trapped in your app without Yidam, I have to:

  • • Scrape the DOM looking for signals (risky, fragile)
  • • Guess based on user input (often vague)
  • • Hope you stuffed enough state into my prompt (you didn't)

With Yidam, I simply listen to synced state and do the right thing."

— Your AI (any LLM)

Dead-Simple Integration

Most apps only need 3–5 syncs to unlock powerful AI.
You already have the state — Yidam just taps in.

1. Install

npm install @yidam/sdk

2. Initialize

import { yidam } from '@yidam/sdk';yidam.init({ apiKey: 'your-key' });

3. Sync Just What Matters

const [screen] = yidam.sync("screen", currentRoute);const [form] = yidam.sync("form", formState);

That's it. No wrappers, no rerenders, no guesswork.

🧠Your AI now sees what your user sees.

Works With Everything

Any LLM: OpenAI, Claude, Gemini, Mistral, local models
Any Framework: React, Vue, Angular, Vanilla JS
Any Use Case: Chat, copilots, form fills, search, automation
BYOK or Managed: Use your keys or plug into ours

Simple, Transparent Pricing

Starter

$99/mo

For MVPs

  • 10,000 syncs/mo
  • All core features
  • Community support

Growth

$299/mo

For production apps

  • 100,000 syncs/mo
  • Priority support
  • Usage analytics

Scale

$999/mo

For growing companies

  • 1,000,000 syncs/mo
  • Dedicated support
  • SLA guarantee

Need more? Contact us for enterprise pricing.

Be among the first to supercharge your AI with real-time context

Frequently Asked Questions

Is this just for chat?

No! Yidam works for any AI integration - chat, copilots, form fills, search, automation. If you're calling an LLM, we make it smarter.

Do I need to change my AI provider?

Not at all. Yidam works with any LLM - OpenAI, Claude, Gemini, local models. Use your existing setup.

How is this different from prompt engineering?

Instead of manually crafting context into every prompt, Yidam automatically injects real-time app state. It's dynamic and always current.

Is my data secure?

Yes. Context data is processed in-memory and never stored. You can also self-host for complete control.

How long does integration take?

Most developers are up and running in under 30 minutes. It's literally 3 lines of code to start.

Ready to Give Your AI Eyes?

Join developers building smarter AI experiences with Yidam.

Questions? Chat with our team (powered by Yidam, of course)