Bring Your Own Compute
We built a rendering platform that costs us nothing to run.
That is not a boast about efficiency — it is the entire pricing architecture. JimFluencer never buys GPU capacity, so it never needs to resell it to you. Every render executes on compute you control.
RunPod Serverless
Cloud rendering, per-second billing
You supply a RunPod API key. Jobs are dispatched to either the JimFluencer managed endpoint or your own custom endpoint ID, and RunPod charges your card directly for the seconds each worker was awake. Workers scale to zero between jobs, so an idle studio bills nothing.
- No hardware, no drivers, no CUDA troubleshooting
- Datacenter A100, H100, or RTX 4090 class GPUs on demand
- Key encrypted with AES-256-GCM before storage
- Connection test before you ever queue a job
Local ComfyUI
Your hardware, zero marginal cost
Already own a 3090, 4090, or better? Enter your local ComfyUI address — a LAN address such as http://127.0.0.1:8188 or a Cloudflare Tunnel hostname for remote access. The Studio interface behaves identically and rendering costs nothing beyond electricity.
- Free unlimited rendering on hardware you already own
- Your model weights and datasets never leave your machine
- Cloudflare Tunnel support for rendering while travelling
- Same workflow normalisation as the cloud path
Real Numbers
What rendering actually costs
Estimates based on RTX 4090 serverless rates at typical step counts. These are provider charges billed to you, not JimFluencer fees.
| Workload | GPU time | Your cost |
|---|---|---|
| High-resolution still (RealVisXL, 30 steps) | ≈2s | ~$0.002 |
| Cinematic reel, 5.4s at 24fps (LTX 2.3, 129 frames) | ≈25s | ~$0.015 |
| Cinematic reel, 10.7s at 24fps (LTX 2.3, 257 frames) | ≈50s | ~$0.030 |
| 30-day content pack (24 images + 6 reels) | ≈3m 45s | ~$0.14 |
| Character LoRA training run (SDXL, 24 images) | ≈18m | ~$0.65 |
Encrypted before storage
Your RunPod key is sealed with AES-256-GCM the moment it is submitted. The plaintext never returns to the browser — the settings panel shows only the final four characters.
Scoped to your account
Because the key is yours, spending limits, usage alerts, and revocation all stay under your control in the RunPod dashboard. Revoke it and rendering stops immediately.
Defensive dispatch
Every worker response is validated before parsing. Gateway timeouts and HTML error pages become structured JSON errors instead of unhandled crashes mid-render.