Meshive GPU Cloud logoMeshive

ComfyUI, running in two minutes

Generative image and video work is spiky — a week of nothing, then a night of batches. Meshive is priced for that shape: start the pod for the batch, stop it when the queue drains.

Start a podSee pricing
An iridescent translucent flower rendered in teal and violet light
A flat paper-cut illustration of flowing wave ribbons in violet, pink and teal
01

Fourteen model packs, not one ComfyUI

FLUX.1, SDXL, Qwen-Image, Z-Image, Krea 2, Wan 2.1 and 2.2, LTX-Video and MiniMax H3, each as its own pack.

Long-exposure streaks of blue and teal light
02

The weights attach at launch

Pick a pack and the model set downloads to managed storage when the pod starts. Nothing is baked into the image and nothing is fetched twice.

A video editor's dark desk with several monitors showing a timeline and color grade
03

Edit the graph, or just call it

Run the pod when you need to change the workflow. The same models sit on Serverless for when you only need the output.

Ready to launch

Templates for this work

Official images with the drivers and the stack already in place.

  • ComfyUI - FLUX.1-dev
  • ComfyUI - Wan 2.2 I2V
  • ComfyUI - Qwen-Image
  • ComfyUI - Z-Image Turbo
  • ComfyUI - SDXL
  • ComfyUI - LTX-Video
Made on Meshive

Rendered on a single RTX 5090

  • A photoreal still life of a cut-glass tumbler of amber liquid beside a linen cloth and a sprig of rosemaryKrea 2
  • A rain-wet night alley strung with paper lanterns, steam rising beside a parked bicycleKrea 2
  • A flat paper-cut poster illustration of layered mountains at sunsetKrea 2
  • MiniMax H3
Hardware

Cards that suit this work

Every Meshive pod gets a whole, non-virtualized GPU. These are the ones we would reach for first.

  • RTX 509032GB VRAM · ConsumerSDXL, FLUX.1 and Wan 2.2 fit in 32GB
  • RTX 409024GB VRAM · ConsumerThe workhorse for most ComfyUI graphs
  • RTX PRO 5000 Blackwell48GB VRAM · WorkstationThe larger FLUX.1 and Wan checkpoints at 48GB
  • RTX PRO 6000 Blackwell96GB VRAM · WorkstationLonger video and batch runs at 96GB
All GPUs and current pricing
A GeForce RTX card standing edge-on against near-black, its edges rimmed in green and blue light
Getting started

Three steps

  1. 1Pick the pack for the model you want - the weights attach when the pod starts.
  2. 2Open ComfyUI in the browser and load the workflow that ships with it.
  3. 3Render. Outputs land on the volume and survive the pod.

Also on Meshive

RTX PRO 5000 · 48GB

3D rendering

L40S · 48GB

Inference serving

City lights seen from low orbit at night, curving along the dark horizon of the Earth

Pick a card and start the pod

Sign up, choose the GPU, and the pod is yours in under two minutes. It bills by the hour and stops when you stop it.

Already a user? Invite friends and earn 15% of their first top-up.

Get started freeSee pricing