What is Capi?

Capi is the memory layer for AI applications. It solves the problem of stateless AI by:

Remembering user interactions and preferences
Retrieving relevant context when needed
Scaling from prototypes to production

Why Use Capi?

Personalization: Build AI that truly understands each user
Cost-Saving: Reduce redundant LLM context tokens by 60-90%
Scalability: Handles from 100 to 100M+ memories
Simplicity: Just two API endpoints to learn

Core Concepts

Memories: Stored user interactions with text and metadata
Users: Unique identifiers for memory isolation
Context: Retrieved memories relevant to the current interaction
Embeddings: Vector representations enabling semantic search

Quickstart Guide (5 Minutes to First Memory)

Go to capi.dev/sign-up
Create your first project
Generate an API key in the dashboard

Step 2: Store Your First Memory

curl -X POST https://capi.dev/api/v1/addMemory \
  -H "Content-Type: application/json" \
  -H "X-CAPI-API-Key: your-api-key-here" \
  -d '{
    "userId": "user-123",
    "text": "User prefers dark mode and speaks French"
  }'

Step 3: Retrieve Context

curl -X POST https://capi.dev/api/v1/retrieveContext \
  -H "Content-Type: application/json" \
  -H "X-CAPI-API-Key: your-api-key-here" \
  -d '{
    "userId": "user-123", 
    "query": "What interface preferences does this user have?",
    "limit": 3
  }'

Step 4: View Usage in Dashboard

Go to your project dashboard
Check the "Usage" tab to see:
- Memory operations
- Storage usage
- API call history

Step 5: Next Steps

Authentication

Getting Your API Key

Log into capi.dev or create an account
Select your project
Navigate to "API Keys"
You should already have a key created for you. if not Click "Generate New Key" to create a new one.

Using the API Key

Include it in the X-CAPI-API-Key header for all requests:

fetch('https://capi.dev/api/v1/addMemory', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'X-CAPI-API-Key': 'your-key-here' // 👈 Required header
  },
  body: JSON.stringify(payload)
});

Security Best Practices

Never expose keys in client-side code
Rotate keys every 90 days
Use environment variables:
```
# .env
CAPI_API_KEY=your-key-here
```
Restrict keys to specific IPs if possible

Pro Tip

Use different API keys for development and production environments.

Full Code Example

// Complete Node.js example
const CAPI_BASE_URL = 'https://capi.dev/api/v1';
const API_KEY = process.env.CAPI_API_KEY;

async function chatWithMemory(userId, message) {
  // 1. Store the user message
  await fetch(`${CAPI_BASE_URL}/addMemory`, {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json',
      'X-CAPI-API-Key': API_KEY
    },
    body: JSON.stringify({
      userId,
      text: `User said: ${message}`
    })
  });

  // 2. Get relevant context
  const context = await fetch(`${CAPI_BASE_URL}/retrieveContext`, {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json', 
      'X-CAPI-API-Key': API_KEY
    },
    body: JSON.stringify({
      userId,
      query: message,
      limit: 3
    })
  }).then(res => res.json());

  // 3. Use context in your LLM call
  const prompt = `Context:\n${context.memories.join('\n')}\n\nUser: ${message}\nAI:`;
  return generateAIResponse(prompt);
}

Next Steps

Join our Discord community

Getting Started