News3 min read

Runware x Comfy Router

Runware is available on Comfy Router from day one. Developers can run supported models on Runware's inference infrastructure through Comfy's unified API, with no separate Runware integration required.

Ally Nicoll
Ally Nicoll
•

TL;DR

Comfy Router lets developers call frontier image, video, 3D, and audio models through ComfyUI's SDK and API key. Runware is a launch inference provider. Set model_provider="runware" on a supported model to route your Comfy jobs to us.

We're excited to announce that Runware is a launch inference provider for Comfy Router, Comfy's new unified API for running frontier image, video, 3D, and audio models through a single integration. Runware is available as a provider from day one.

What Comfy Router does

Comfy Router brings generative media models into one API, with a shared way to submit jobs and track results. Developers can add supported models without building a new integration each time.

Runware is available on Comfy Router for models including GPT Image 2, GPT Image 2.5 Flare and Sunburst, Nano Banana 2 and Pro, and Seedance 2.0 and 2.5. That means developers can access those models on Runware from the Comfy SDK they already use.

Why run Comfy Router on Runware

Runware is built to make generative media inference fast and efficient at scale. Our Sonic Inference Engine handles the work around each request, including routing, scheduling, and managing capacity as demand changes. Through Comfy Router, developers can choose Runware for supported models without building a separate integration.

For developers, the integration is simple: choose a model supported on Runware and set model_provider="runware". The request still goes through Comfy Router, so there's no additional Runware account, API key, or job lifecycle to manage.

Getting started

Comfy Router launched with Runware already wired in, so there's no separate signup or setup on our end. Point the provider parameter at runware and everything else — the model string, the arguments, the job lifecycle — works exactly as Comfy Router's own docs describe.

curl -X POST "https://api.comfy.org/v2/models/openai/gpt-image-2?\
model_provider=runware" \
-H "X-API-Key: $COMFY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $(uuidgen)" \
-d '{"prompt": "aerial view of a neon coral reef at dusk"}'

Since Comfy Router is brand new, their quickstart docs are the best place to get the full picture: the SDK, authentication, and job queuing. The runware provider parameter gives your Comfy Router integration access to Runware's inference infrastructure.

Written by

Ally Nicoll

Developer Advocate at Runware

Ally Nicoll is a Developer Advocate at Runware, where he helps developers and creators understand and build with Runware's AI infrastructure. He focuses on developer education, product communication, model launches, and practical AI workflows, translating technical capabilities into clear, useful guidance.

More posts by Ally Nicoll
Articles|

Run the fastest, lowest-cost generative AI API.

Start with free test credits.

Get started now

Keep reading

Runware expands engineering leadership team with hires from NVIDIA and OpenAI
News•Nov 5, 2026•6 min read

Engineering leaders from NVIDIA, OpenAI, and NetApp join Runware

Senior hires from NVIDIA, OpenAI, and NetApp take over Runware's infrastructure and security, with further senior additions from Amazon Web Services and Juniper Networks.

Aerial render of the Runware factory floor, where Sonic Pod AI data centers are assembled on a production line
News•Oct 30, 2026•5 min read

Runware unveils AI factory to build modular AI data centers just like cars

Runware's new European factory will build complete, AI-native data centers on an assembly line: 1MW and 2MW Sonic Pods, up to 30MW a day, 10GW a year, from Q1 2027.