Bring your own GPU (0 credits)

The zero-credit path. Connect the GPU in your PC — or a cloud GPU you rent yourself on RunPod or Vast — and image, video, and voice generation all run on that hardware. Costs stay on your machine or your rental account; not a single owntent credit is spent.

1. 1Local PC (easiest): with an RTX 3090/4090-class GPU, install ComfyUI and connect its address (default http://127.0.0.1:8188) in the project's ComfyUI studio tab.
2. 2Writing LLM locally too: run an OpenAI-compatible server such as Ollama and point the AI backend in Settings at it — scripts and dialogue also generate for 0 credits.
3. 3Cloud rental (no hardware): rent a GPU instance yourself on RunPod, Vast.ai, or similar. Hourly charges bill to your own account; owntent is not involved.
4. 4In the rented instance's terminal, run the one-line bootstrap: curl -fsSL https://owntent.com/gpu/bootstrap.sh | bash — it installs ComfyUI plus a default video model (LTX) and prints the access address.
5. 5Paste the printed public address into the ComfyUI studio tab and confirm the connection.
6. 6Want several cuts at once? Launch several instances. Video generation runs one cut per GPU at a time, so parallelism scales with instance count. When the batch is done, shut the instances down — you pay for every hour they stay up.

ComfyUI workflows (open-weight models such as LTX video) and OpenAI-compatible local LLMs such as Ollama. See the ComfyUI studio doc for connection and model management.