Updated 2026-10-07
RebusFarm, GarageFarm.NET and Fox Renderfarm are commercial render farms for studios: you submit from a desktop client or plugin, they spread a job over tens or hundreds of nodes, and you pay per node-hour, GHz-hour or OctaneBench-hour with volume discounts and $25 trial credits. JANCTION Render is a render service for AI agents: Claude, ChatGPT, Codex or Cursor call it as an MCP tool or an HTTP API, send a .blend, a bpy script or a 3D file, see a tiled preview in seconds, and get frames or an MP4 from one datacenter GPU, free during the beta (10 GPU-minutes per key per day, 240 frames at 1080p per job) and metered per GPU-second after. For a feature-length sequence on hundreds of nodes with studio support, use the big farms; for a scene an agent wrote ten seconds ago, a product turntable, or a script with no Blender installed, use JANCTION Render.
| JANCTION Render | GarageFarm.NET | Fox Renderfarm | RebusFarm | |
|---|---|---|---|---|
| Who presses the button | an AI agent (MCP tool), a script (HTTP API), you (CLI, Blender add-on) | you, from renderBeamer (desktop client) and DCC plugins | you, online submission or desktop client | you, from their desktop app and plugins |
| Inputs | .blend, bpy script, glTF / FBX / USD / OBJ / STL / PLY / Alembic | Blender and other DCC projects, packaged by the client | Blender and many other renderers' projects | Blender and other DCCs (see their software list) |
| Hardware | one NVIDIA RTX PRO 6000 (96 GB) during the beta; frames split into chunks | up to 15 / 30 / 60 GPU nodes or 100 / 150 / 300 CPU nodes per priority tier | CPU and GPU node pools | CPU and GPU pools |
| Price unit | GPU-second (planned 0.1 JPY); free beta today | GPU: $0.004 / $0.006 / $0.012 per OctaneBench-hour; CPU: $0.024 / $0.036 / $0.072 per GHz-hour (low / medium / high priority) | CPU: $1.224 per node-hour, down to $0.734 with recharge tiers; GPU rates behind a toggle not readable on the page | pricing page returned HTTP 404 on 2026-10-07 |
| Free tier | 10 GPU-minutes per key per day, no sign-up | $25 trial credits (10 nodes, 30 jobs) | $25 free trial | not read |
| Preview before the final | 1 to 4 frames tiled in one image in a few seconds, returned to the agent | test renders via the client | test renders via the client | not read |
| Estimate | GPU seconds from measured per-frame times before the job | cost estimator on the site | cost estimator on the site | not read |
| Agent access | remote MCP (OAuth), stdio MCP, OpenAPI, webhooks | no public agent API found | no public agent API found | no public agent API found |
| Support | email, next business day (planned Pro plan); beta as is | email; 24/7 on medium and high priority | 24/7 per their site | per their site |
Units are not comparable (an OctaneBench-hour measures GPU throughput, a GHz-hour CPU throughput, a node-hour a whole machine), so the honest comparison is one of your own scenes on each. Measured on JANCTION Render: a 1080p Cycles still in 5 to 10 GPU-seconds, a 24-frame 720p turntable in 58 GPU-seconds (costs).
They combine: iterate on JANCTION Render, then send a long final sequence to a many-node farm. Other comparisons: BlendSwap, SheepIt, render farm APIs, all.
| from | how |
|---|---|
| A file on this computer (.blend, 3D file) | Drop it at /upload to get a 12-hour link (no sign-up), then tell the AI "Render this file: <link>". From inside Blender use the add-on; from a terminal the CLI uploads directly |
| 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 |
| Gemini CLI / Antigravity CLI | gemini extensions install https://github.com/JasmyLab-JANCTION/janction-render |
| Grok / Perplexity / Le Chat | Add a custom (remote MCP) connector with https://render.janction.jp/mcp in each app's Connectors settings (Grok: New Connector → Custom; Perplexity: Pro / Max / Enterprise; Le Chat: workspace admin) |
| Windsurf, Cline, Goose, 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 |
Not comparable by unit. At the planned 0.1 JPY per GPU-second a 24-frame 720p turntable is about 6 JPY and a 1080p still under 1 JPY; the big farms bill per node-hour or benchmark-hour with volume discounts. Run your own scene on each to know.
Not during the beta: 240 frames and 1080p per job, one GPU. That is big-farm territory today.
No. Blender with Cycles (and EEVEE) only. The large farms support many renderers through their plugins.
They are driven from a desktop client, so there is no tool an agent can call, no preview it can look at, and no estimate it can quote. JANCTION Render is built as a tool: MCP, HTTP API, previews and estimates in the response.
JANCTION Render v0.4.20 · JasmyLab Inc. · Updated 2026-10-07 · Tested with Blender 5.0, JANCTION Render 0.4.20 · 日本語