Kling 3.0 Turbo · Text to Video

Six widgets. That's the whole node. Kling 3.0 Turbo Text to Video is the minimal way to get a Kling clip out of a ComfyUI graph: prompt, aspect ratio, resolution, length, and out comes an MP4 with sound attached. It runs from 16.8 credits per second, more than the non-Turbo Kling 3.0 text node costs for a lot more control - so treat this as the express lane, not the budget option.

How it works

The node turns your prompt and settings into a request, gets a quote from PopNow, submits the task with an idempotency key, polls with progress text printed on the node itself, then downloads the finished clip and returns it as a VIDEO. Kuaishou's model runs on PopNow's servers. No local GPU, no sampler settings, no seed hunting - and per the node's own description, "sound on every clip", so there's no silent mode to switch to.

The inputs

prompt - scene, action, camera. With this node always generating audio, add spoken lines in quotes and name the sounds you want, in their own sentences. "A woman walks through a wet market, camera tracking, she says 'we're late'" gets you both the picture and the line.

aspect_ratio - 16:9, 9:16, or 1:1. Vertical if it's going to a phone.

resolution - 720p or 1080p. There's no 4K on the Turbo node.

duration_seconds - whole seconds, 3 to 15. At base rate the ceiling is 252 credits (about a quarter), and it's billed per second, so this is your main cost dial.

That's it. No guidance dial, no negative prompt box, no shot list, no closing frame. If you want any of those, you want the plain Kling 3.0 text-to-video node instead.

Outputs

video (VIDEO) and task id (STRING). The video is a standard ComfyUI value - Save Video, or into another PopNow node to extend or edit. The task id identifies the run in the API console, and it's the only thing you can quote to support when something goes sideways.

Install

Manager → search PopNow → Install → restart. Or:

cd ComfyUI/custom_nodes
git clone https://github.com/popnow-ai/popnow-comfyui

Then create a key at popnow.ai/api/console/keys, paste it into Settings > PopNow, or export POPNOW_API_KEY - the env var wins if both are set. Restart again afterwards: the node list is resolved from your account's catalog at startup. The pack installs no models and declares no dependencies; it only uses what ComfyUI already ships, which is why this works on a laptop, a Mac, or a CPU-only box. Delete any leftover custom_nodes/popnow_api.py.

Watch-outs

The pricing is the headline. The Turbo badge sits at 16.8 cr/s where the fuller Kling 3.0 node sits at 10.08, so a 15-second take is 252 credits vs 151 credits for more controls. Pay the premium because you want faster turnaround, not because "Turbo" sounds cheaper.

Cancellation doesn't cancel. Interrupt in ComfyUI stops the wait; the task keeps rendering on PopNow, is charged if it succeeds, and the output lands in the API console. If you kill a run, check the console before queueing again.

Set Settings > PopNow > Max credits per run and the node refuses anything quoted above it, before credits are held. That is the single most useful setting in the pack while you're learning a model's cost profile.

Codes: No PopNow API key is set / 401 (no key or revoked), 40201 (out of credits), 40004 (unservable combination - try a shorter duration), 40310 (verify your email), content_rejected and generation_failed (both refunded). Anything with a request_id in it is a support ticket with the receipt attached.

And since Kling 3.0 was a genuinely contested release - some people found it a step back from 2.6, others thought it was the best thing available - compare it on your own prompts before you build a pipeline on the Turbo speed advantage.