Updated 2026-10-01
Send the scene (a .blend, a bpy script or a 3D file) with a frame range and you get the frames rendered on an NVIDIA RTX PRO 6000 and joined into a video: janction-render render scene.blend --frames 1-96 --size 1280x720 --samples 64 --wait, or render_final(frame_start=1, frame_end=96, output='mp4') from an agent. Measured: about 2.4 GPU seconds per 720p frame at 64 samples on Cycles for a simple scene (24 frames = 59 s); EEVEE is about half per frame for a few dozen frames or more. The free beta renders up to 240 frames per job (split longer animations) at up to 1080p; outputs are mp4 (H.264), webm (VP9), prores (422 HQ), gif, webp or png / exr frames. Preview 4 frames first to check camera and motion.
janction-render preview scene.blend --frames 1-96 # 4 frames spread over the range, one sheet janction-render render scene.blend --frames 1-96 --size 1280x720 --samples 64 --wait # output.mp4 janction-render render scene.blend --frames 1-96 --engine eevee --output webm --wait # cheaper per frame janction-render render scene.blend --frames 1-96 --output prores --wait # for editing janction-render render model.glb --frames 1-48 --orbit --environment studio --wait # turntable of a model
From an agent the same is render_preview(frames='1-96') then render_final(frame_start=1, frame_end=96, width=1280, height=720, samples=64, output='mp4'); render_estimate first tells you the time and whether it fits today's quota.
| job | GPU time | note |
|---|---|---|
| preview, 4 frames (720p split in 4 tiles, 16 samples, denoised) | about 4 s | round trip about 8 s; the image comes back inline |
| final, 24 frames at 1280x720, 64 samples → MP4 | about 60 s | measured 2026-09-30 / 10-01 (59 s and 58 s) |
| final, one 1920x1080 frame at 128 samples | about 11 s | the estimator's baseline; heavy scenes take longer |
render_estimate gives the number before rendering; once a scene has been rendered once, the estimate uses that scene's own measurement.
Frames are split into chunks and rendered in order; while a job runs, render_status shows frames done and the ETA. The benchmarks page has the per-frame table for several resolutions and sample counts.
Cycles (default) is the path tracer: accurate light, slower. engine='eevee' is Blender's real-time rasterizer: about half the per-frame cost, but every job pays about 7 seconds of shader compilation, so it only pays off for a few dozen frames or more; previews stay on Cycles. HDRI presets work on both.
| output | what you get | use it for |
|---|---|---|
| mp4 | H.264, fps as given | sharing, chat, X |
| webm | VP9 (alpha with transparent=true) | web, overlays |
| prores | ProRes 422 HQ .mov | editing |
| gif / webp | animated, at most 960 px wide | chat and README embeds |
| png / exr | one file per frame (exr = 16-bit linear) | compositing, your own encode |
Free beta (as of 2026-10-01): 10 GPU-minutes per key per day (20 per network), final renders up to 240 frames at 1080p. No sign-up: connecting creates a key. Inputs and results are deleted 24 hours after last use and never used for training. Paid plans will be announced on this site and in llms.txt before they start.
| from | how |
|---|---|
| Claude.ai (web / desktop / mobile) | Settings → Connectors → Add custom connector → https://render.janction.jp/mcp. "Connect" opens a page that creates a free key |
| ChatGPT | Settings → Connectors → Advanced → Developer mode → Create → https://render.janction.jp/mcp (OAuth) |
| Claude Code | claude mcp add --transport http janction-render https://render.janction.jp/mcp, then /mcp to authenticate |
| Codex | codex mcp add janction-render --url https://render.janction.jp/mcp |
| Cursor / VS Code | Add to Cursor Add to VS Code The button opens the editor; confirm and it is added. Manual: Streamable HTTP at https://render.janction.jp/mcp |
| Other MCP clients | Streamable HTTP at https://render.janction.jp/mcp; OAuth or Authorization: Bearer <api key> |
| Terminal (stdio, sends local files) | claude mcp add janction-render -e JANCTION_RENDER_SERVER=https://render.janction.jp -- uvx --from janction-render janction-render-mcp |
| Inside Blender (for people) | The add-on; extension repository https://render.janction.jp/extensions/index.json. Maya / Houdini / Cinema 4D: submit tools |
Render it in jobs of up to 240 frames (1-240, 241-480, ...) and join the videos, or request png frames and encode once. The cap is for the free beta.
Yes: fps (default 24) is used when the frames are joined into mp4 / webm / prores.
A preview renders 4 frames spread over the range in a few seconds, so camera, motion and lighting mistakes cost seconds instead of minutes.
Everything in the .blend that Blender renders in the background works, including motion blur and compositor nodes; the service only sets resolution, samples, denoising and the output format.
JANCTION Render v0.4.8 · JasmyLab Inc. · 日本語