Official JANCTION Render: https://render.janction.jp
Short answer: keep editing on your machine and send the render to a GPU somewhere else. JANCTION Render is a cloud GPU render farm for Blender that takes a .blend or a bpy script and returns frames or an MP4. On it, Blender's own Classroom demo scene rendered at 1920x1080 with 128 Cycles samples and denoising in 8.93 GPU-seconds per frame, which is 0.89 JPY per frame, and the Barcelona Pavillion demo in 10.55 GPU-seconds (1.05 JPY). A simple 24-frame 1280x720 animation at 64 samples took about 59 GPU-seconds. Each new key gets 500 JPY of GPU time free (5,000 GPU-seconds, 14 days, no card).
If Cycles is slow on your MacBook
Try the free fixes on the Mac first. Cycles can use the Apple GPU through Metal (Preferences, System, Cycles Render Devices). EEVEE is much faster for drafts. Fewer samples with denoising often look the same as many samples without it.
When the final still takes hours, move only the final. You keep working while it renders, and you pay per GPU-second instead of keeping a machine busy.
If your laptop has no NVIDIA GPU
Send the file. There are three ways in, depending on where you work.
From a chat app such as Claude.ai or ChatGPT: add https://render.janction.jp/mcp as a connector, drop the .blend at https://render.janction.jp/upload (the link is valid for 12 hours), paste the link and ask for a preview, then the animation.
From a terminal:
pip install janction-render
janction-render preview scene.blend
janction-render render scene.blend --frames 1-48 --size 1280x720 --output mp4 --wait
The first command creates a free key. The preview comes back in seconds, and the render command waits for the MP4 and downloads it.
From a headless server or a script
blender -b scene.blend -a works on a server only if Blender is installed there and, for reasonable speed, a GPU that Cycles supports is attached. Without either, the same job goes to the HTTP API or the Python client:
from janction_render.client import Client
c = Client() # or Client(api_key="jr_...")
scene = c.upload("scene.py") # a bpy script, a .blend, or a .glb / .fbx / .usd / .obj file
job = c.submit(scene["scene_id"], kind="final", frame_start=1, frame_end=48, output="mp4")
job = c.wait(job["job_id"], timeout=1800)
print(job["status"], c.download(job["job_id"], "out", only="mp4"))
A bpy script only builds the scene, the camera and the frame range; the service sets resolution, samples and the GPU. Blender 5.0 and 5.2 are available.
Is it cheap?
It costs 0.1 JPY per GPU-second, and the GPU seconds include Blender starting and loading the scene. From the measurements above, a 1080p frame of a demo-scene-sized file is about 1 JPY, and a 24-frame 720p clip of a simple scene is about 6 JPY. Previews are free up to 2 GPU-minutes a day after the welcome credit, so checking a scene usually costs nothing. Each key has a spending cap per job and per day, and a job that needs more credit than the key has is refused before it runs.
The numbers and the scripts that produced them: https://render.janction.jp/benchmarks
Limits
A job goes up to 240 frames at 1080p, so a longer animation is split into several jobs. Inputs and results are deleted 24 hours after last use (7 days for keys that have topped up). There are two GPUs, so a job waits in a queue when both are busy, and the status page says so.
More: https://render.janction.jp/render-blender-without-gpu, https://render.janction.jp/render-blender-animation-online and https://render.janction.jp/blender-render-cost. Prices: https://render.janction.jp/pricing
Official JANCTION Render: https://render.janction.jp. Code and client: https://github.com/JasmyLab-JANCTION/janction-render
Top comments (0)