Migrating from OpenAI to OpenModex in Under 5 Minutes

If you are currently using the OpenAI API directly, you are locked into a single provider. Migrating to OpenModex takes fewer than five minutes and immediately gives you access to 1,600+ models across OpenAI, Anthropic, Google, Mistral, and dozens more -- all through the same API format you already know.

Here is exactly how to do it.

Step 1: Install the OpenModex SDK

Replace your OpenAI SDK with the OpenModex client. The API surface is intentionally compatible, so your existing code patterns carry over directly.

// Before
import OpenAI from "openai";
const client = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });

// After
import { OpenModex } from "@openmodex/sdk";
const client = new OpenModex({ apiKey: process.env.OPENMODEX_API_KEY });

That is the entire SDK swap. If you are using Python, the process is identical:

# Before
from openai import OpenAI
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# After
from openmodex import OpenModex
client = OpenModex(api_key=os.environ["OPENMODEX_API_KEY"])

Step 2: Update Your API Calls

OpenModex uses the same chat completions format. The only difference is that you can now specify models from any provider using a unified naming scheme.

const response = await client.chat.completions.create({
  model: "openai/gpt-4o",          // Same model, explicitly namespaced
  messages: [{ role: "user", content: "Explain quantum computing" }],
  temperature: 0.7,
});

Want to try Anthropic's Claude instead? Just change the model string:

const response = await client.chat.completions.create({
  model: "anthropic/claude-sonnet-4-5-20250929",   // Switch providers instantly
  messages: [{ role: "user", content: "Explain quantum computing" }],
  temperature: 0.7,
});

No new SDK, no new authentication, no new response parsing. Same code, different model.

Step 3: Enable Smart Routing (Optional)

Instead of hardcoding a model, let OpenModex automatically choose the best one for each request based on your priorities.

const response = await client.chat.completions.create({
  model: "auto",                    // Let smart routing decide
  messages: [{ role: "user", content: "Summarize this document..." }],
  routing: {
    strategy: "cost-optimized",     // or "quality" or "latency"
  },
});

Smart routing analyzes each request and selects the optimal model in real time. Our benchmarks show a 37% average cost reduction with no measurable quality loss.

Step 4: Bring Your Own Keys (Optional)

If you already have API keys with OpenAI or other providers, you can use them through OpenModex and skip our markup entirely. Add your keys in the dashboard under Settings > Provider Keys, then enable BYOK mode:

const client = new OpenModex({
  apiKey: process.env.OPENMODEX_API_KEY,
  byok: true,   // Use your own provider keys
});

You still get smart routing, caching, analytics, and failover -- but traffic goes through your own provider accounts.

Step 5: Verify and Monitor

After migrating, check the OpenModex dashboard to verify requests are flowing correctly. You will immediately see per-request cost breakdowns, latency metrics, and cache hit rates.

Run your existing test suite -- since the API format is compatible, your tests should pass without modification. If you encounter any edge cases, our migration guide at docs.openmodex.com/migrate covers provider-specific nuances.

What You Get After Migration

With a five-minute migration, you unlock capabilities that would take months to build yourself: automatic failover across providers when one goes down, semantic caching that eliminates redundant API calls, real-time cost analytics, and access to every major AI model through one integration.

The best part is that switching is risk-free. Keep your existing OpenAI keys as a backup, try OpenModex in staging first, and roll it out when you are confident. Most teams see measurable cost savings within the first week.