Migrate from fal.ai to Runflow in 10 Minutes

Moving from fal.ai to Runflow takes about 10 minutes. The API surface is intentionally similar—we wanted migrations to be painless because that’s how you earn trust.

Here’s the step-by-step.

Step 1: Install the SDK

npm install @runflow/sdk
# or
bun add @runflow/sdk

Step 2: Swap the Import

// Before (fal.ai)
import { fal } from "@fal-ai/client";

// After (Runflow)
import { runflow } from "@runflow/sdk";

Step 3: Update the Endpoint

The request format is nearly identical. Key differences:

Step 4: Add Quality Scoring

This is the main upgrade. Every Runflow inference job can include automatic quality scoring:

const result = await runflow.run("flux-dev", {
  prompt: "Professional corporate headshot...",
  score: true,  // Enable Sentinel scoring
  scoreNiche: "corporate-headshot",
});

console.log(result.image);  // Generated image URL
console.log(result.score);  // 0-100 quality score

See the Score Image API docs for the full scoring API reference.

Step 5: Deploy and Verify

Run your existing test suite against the new SDK. The response format is compatible—image URLs, metadata, and timing fields are in the same structure.

Most teams complete the migration in a single PR. The hardest part is usually updating environment variables in CI/CD.

Full API documentation is at runflow.io/docs. If you hit any issues during migration, reach out—we’ll pair with your team to get it done.