SUNDAY, NOVEMBER 15, 2026
Render · Blender on RTX PRO 6000 · Tokyo
Front page / Status
Status
Now: operating. 3 GPUs (5 workers) can take jobs.
JANCTION Render is a cloud GPU render farm for Blender that AI agents call as an MCP server or HTTP API. Connect: MCP server https://render.janction.jp/mcp (Claude, ChatGPT, Codex, Cursor and others; no sign-up). Try it in a browser: /start. Each new key starts with 100 JPY of free GPU time (no card) and 400 JPY more after saving a card that is not charged, up to 500 JPY (14 days); after that, previews stay free up to 2 GPU-minutes a day and finals cost 0.1 JPY per GPU-second. Pricing: /pricing.
| Queued chunks | 0 |
|---|---|
| Running chunks | 0 |
| Expected wait | 0 s |
Workers
| name | GPU | state |
|---|---|---|
| GPU worker 1 | NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | online |
| GPU worker 2 | NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | online |
| GPU worker 3 | NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | online |
| GPU worker 4 | NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | online |
| GPU worker 5 | NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition | online |
Last 7 days
| Renders completed | 685 |
|---|---|
| Success rate | 99% |
| Median queue wait | 0 s |
| Median submit-to-result | 9 s |
| Uptime (checked from outside every 5 minutes) | 100% (2,016 of 2,016 checks) |
| End-to-end check (hourly preview through the MCP) | 100% (161 of 161) |
The success rate is the share of external previews and finals that finished in the last 7 days (done and failed; canceled jobs are not counted) that ended in done. Failures on JANCTION's side are counted; failures caused by what was sent, such as an error in the script or a missing file, and the operator's own tests are not. Uptime is measured, not a guarantee: the share of checks where /v1/health through Cloudflare and the entry points AI apps use (/mcp and the OAuth metadata) all answered correctly. JSON: uptime in /status.json.
Usage (external only)
| renders completed | success rate | GPU minutes | median round trip (s) | apps | |
|---|---|---|---|---|---|
| last 24 h | 115 | 100% | 234.5 | 12 | Claude Code 39, API (own agents and apps) 37, Codex 30, CLI 9 |
| last 7 days | 685 | 99% | 916.5 | 9 | API (own agents and apps) 512, Codex 44, Claude Code 39, Claude 34, ChatGPT 28, CLI 15, MCP (other apps) 9, Grok 4 |
| last 30 days | 1984 | 99% | 2915.9 | 8 | API (own agents and apps) 1382, Claude 123, Cursor 102, MCP (other apps) 100, Codex 67, ChatGPT 57, Claude Code 46, CLI 41, Sesame 17, Notion 14 |
The operator's own tests, benchmarks and monitoring, and testers' keys, are not counted. The success rate counts failures on JANCTION's side; failures caused by what was sent, such as an error in the script or a missing file, are left out. Since launch (2026-09-30): 2,523 renders, 3,699 GPU-minutes. JSON: /v1/stats (refreshed every 5 minutes).
Available right now
Blender 3.6 / 4.2 / 4.5 / 5.0 / 5.2 / 5.3 · environment presets night, overcast, studio, sunset · inputs .abc .blend .fbx .glb .gltf .obj .ply .py .stl .usd .usda .usdc .usdz
This page refreshes every 30 seconds. Machine-readable: /v1/health (JSON). Agents see the same through render_info.