01 / implementation

From zero to production in under five minutes.

Install the SDK, configure your environment, and ship your first AI-powered feature before your next standup.

what eichler ai surfaces

Intelligent mapping for architecturally significant spaces.

Eichler AI is built around one primary tool: a spatial intelligence engine that reads any floor plan — photograph, sketch, or PDF — and surfaces a structured 3D data model of the space. It was designed specifically for the documentation and preservation of architecturally significant properties, starting with the mid-century Eichler homes of the San Francisco Bay Area. The same engine works on any complex or historically important structure where precision documentation matters.

Floor Plan MatrixAI Sightline Mapping3D Track MapVault GeometryPreservation Documentation

Step 01

Install the SDK

Available on npm, PyPI, pkg.go.dev, and crates.io. Pick your language and install.

terminal

npm install eichler-ai

Step 02

Set your API key

Add your key to your environment. Never commit it to source control.

terminal

EICHLER_KEY=ek_live_xxxxxxxxxxxxxxxxxxxx

Step 03

Initialize the client

Create a single client instance and reuse it across your application.

step-3.ts

import { EichlerAI } from 'eichler-ai';
const client = new EichlerAI({
  apiKey: process.env.EICHLER_KEY,
  model: 'eichler-3-turbo',
});

Step 04

Make your first call

Stream a completion and handle the response in real time.

step-4.ts

const stream = await client.complete({
  prompt: 'Explain quantum entanglement simply.',
  maxTokens: 512,
  stream: true,
});
for await (const chunk of stream) {
  process.stdout.write(chunk.text);
}

Environment configuration

Eichler AI adapts to your deployment target. Set the right variables for each environment.

Development

.env

EICHLER_KEY=ek_test_xxxx
EICHLER_LOG_LEVEL=debug
EICHLER_BASE_URL=https://api.eichler.ai/v2

Production

.env

EICHLER_KEY=ek_live_xxxx
EICHLER_LOG_LEVEL=error
EICHLER_TIMEOUT_MS=30000
EICHLER_RETRY_LIMIT=3

Deployment patterns

Singleton pattern

Instantiate once at module level and import the instance wherever you need it. Avoids connection overhead on every request.

Edge-compatible

The SDK ships a fetch-based transport with no Node.js-specific dependencies. Deploy to Cloudflare Workers, Vercel Edge, or Deno without modification.

Retry & backoff

Automatic exponential backoff on 429 and 5xx responses. Configure EICHLER_RETRY_LIMIT to tune for your SLA requirements.

OpenTelemetry traces

Pass a tracer instance to the client constructor and every request emits spans to your existing observability stack automatically.

Ready to go deeper?

Explore the full API reference and authentication patterns.

API Integration