Deploy Workflows as APIs

Transform your ComfyUI workflows into production services with three deployment options:
Use CaseBest SurfaceNotes
Quick reviews, creative iterationPlayground UIGuardrails + presets; zero code.
Product / tool integrationAPI or SDKStable contract, versioned schema, auth.
Automation, batch, agentsSDK + Scaling controlsQueues, concurrency, budget caps.

Core Capabilities

  • Configure deployments with runtime controls and access permissions
  • Launch interactive playground UIs for non-technical users
  • Integrate workflows programmatically via SDKs or REST APIs
  • Auto-scale GPUs with cost controls and performance monitoring

Audience Quick-Nav

For Developers

Need: Auth, endpoints, payload formats, SDK examplesQuick Start: JS/TS SDK or Python SDK

For AI & Product Leads

Need: Governance, safe defaults, usage limits, team accessQuick Start: Deployment Configuration

For Leadership

Need: Risk controls, scaling behavior, integration patternsQuick Start: Security & Governance

Mental Model

Every deployed workflow in FlowScale AI follows this progression: Workflow Artifact (your ComfyUI graph + assets)
Deployment Configuration (runtime, schema, permissions)
Service Surfaces (API endpoint, Playground UI, SDK bindings)
Scale & Governance (metrics, scaling, budgets)

Quick Start (5 Minutes)

1

Deploy a Workflow

Pick a saved workflow and click Deploy → Select runtime + access scope → Publish v1
2

Test with Playground

Generate Playground Link and test parameter variations with guardrails
3

Call from Code

Copy the API Endpoint → paste into SDK quick start examples below
4

Monitor Runs

Open Runs tab to see generation and outputs real-time

SDK Examples

import { FlowscaleAPI } from 'flowscale';

const flowscale = new FlowscaleAPI({
  apiKey: 'your-api-key',
  baseUrl: 'your-api-url',
  allowDangerouslyExposeApiKey: true // Required acknowledgment
});

const workflowId = "your-workflow-id"; // UUID format
const inputs = {
  prompt: 'A serene mountain landscape at sunset',
  width: 1024,
  height: 1024
}

const result = await flowscale.executeWorkflowAsync(workflowId, inputs);

console.log('Workflow Result:', result);

Service Surfaces Explained

API Endpoints

When to use: Product integration, backend automation, webhook triggers Key features:
  • RESTful HTTP interface with auto-generated OpenAPI docs
  • Strongly-typed input validation and structured outputs
  • Authentication via API keys with fine-grained permissions
  • Async job support for long-running workflows

Playground UIs

When to use: Creative review cycles, stakeholder demos, non-technical team access Key features:
  • Auto-generated interfaces based on workflow parameters
  • Parameter presets and validation guardrails
  • Shareable links with access controls

SDK Integration

When to use: Application development, batch processing, programmatic automation Key features:
  • Type-safe clients for JavaScript/TypeScript and Python
  • Built-in retry logic and error handling
  • Streaming support for real-time outputs
  • Environment-specific best practices (frontend vs. server)

Enterprise Features

  • Cost Controls: Budget caps and spending alerts
  • Brand Safety: Parameter restrictions and model locking
  • Security: HTTPS, audit logging, SOC 2 compliance

Deployment Patterns by Use Case

Pattern: Playground UI with parameter presetsConfiguration:
  • Restricted parameter ranges for brand safety
  • Low-cost GPU tiers for occasional use
Example: Creative team reviews ad variations before campaign launch

Next Steps

Integration

Configuration

Documentation