Quick Start

Get started with Aetherfy in under 5 minutes. Build your first AI agent memory system.

Getting your first API key

Web-first (recommended)

  1. Sign up with email or GitHub.
  2. On first login, the welcome modal reveals your default API key's plaintext.
  3. Copy it — you will see it once.
  4. Use it from any SDK, CLI, or HTTP client.

You will see your key once. The plaintext is never stored on our side after that single reveal — only its hash. If you lose it, head to the API Keys page and create a replacement.

SDK / CLI-first (BYOK)

Already comfortable in a terminal? Once you have your key from the dashboard, drop it into AETHERFY_API_KEY and the SDKs pick it up automatically:

# Python SDK — picks up AETHERFY_API_KEY automatically
import os
os.environ["AETHERFY_API_KEY"] = "afy_live_your_api_key_here"

from aetherfy_vectors import AetherfyVectorsClient
client = AetherfyVectorsClient()  # env var → no kwarg needed
// JS / TypeScript SDK — same env-var convention
// process.env.AETHERFY_API_KEY = 'afy_live_your_api_key_here'

import { AetherfyVectorsClient } from 'aetherfy-vectors';
const client = new AetherfyVectorsClient();  // env var → no kwarg needed

The CLI is the same convention — afy login --api-key … persists the key locally.

A note on revocation:the auto-provisioned default key can be revoked like any other. Revoking it doesn't lose dashboard editor access — the editor uses an internal token — but it does invalidate the key for SDK and CLI use, so create a replacement before revoking if you're using it outside the dashboard.

1

Install the SDK

Add Aetherfy to your project

npm install aetherfy-vectors
2

Initialize the client

Get your API key from the dashboard

import { AetherfyVectorsClient, DistanceMetric } from 'aetherfy-vectors'

const client = new AetherfyVectorsClient({
  apiKey: process.env.AETHERFY_API_KEY
})
3

Create a collection

Set up a collection for your vectors

await client.createCollection('agent-memory', {
  size: 1536,  // OpenAI ada-002 dimensions
  distance: DistanceMetric.COSINE
})
4

Insert vectors

Add vectors with payload data to your collection

await client.upsert('agent-memory', [
  {
    id: 1,  // unsigned integer, or a UUID string
    vector: embeddings,  // Your vector embeddings
    payload: {
      agent_id: 'assistant',
      content: 'User prefers morning meetings',
      timestamp: Date.now()
    }
  }
])
5

Search vectors

Find similar vectors with filtering

const results = await client.search('agent-memory', queryEmbedding, {
  limit: 5,
  withPayload: true,
  queryFilter: {
    must: [{ key: 'agent_id', match: { value: 'assistant' } }]
  }
})

console.log(results) // Returns from the nearest region

What's Next?

Explore the full SDK documentation
Monitor your usage in the dashboardRead the Limits reference (RPM and per-request caps)View pricing and calculate costs