JANCTION Render
A cloud-based Blender rendering service built on the MCP protocol. It encapsulates tools for scene upload, job submission, status polling, and download of finished renders, for use by LLM agents.
Timeline (last 90 days)
Official measurement: calling render_preview for the example scenario with four lighting presets, under the default 5-second timeout it threw McpError: Timed out after 5.1 seconds; after changing to a 120-second timeout the same task returned a result in 7.3 seconds, actually consuming 4.03 GPU seconds.
JANCTION Render provides a remote MCP Server endpoint https://render.janction.jp/mcp, authenticated with a Bearer Token, obtained by calling curl -s -X POST https://render.janction.jp/v1/keys, with no registration required.
Through the JANCTION Render MCP Server, the Agent can see 12 tools, with a typical call chain of scene_info → render_preview → render_estimate → render_final → render_status → render_download.
JANCTION Render gives each new key 500 yen of GPU time (5000 GPU seconds), valid for 14 days, with no card binding required; afterwards preview tasks get 2 GPU minutes free per day, and other GPU time is billed at 0.1 yen per GPU second.
Official measurement: a 24-frame 720p MP4 at 64 samples consumes about 58 GPU seconds, equivalent to about 6 yen.
When the JANCTION Render quota is exhausted, the task is rejected before running, the HTTP API returns 402 with a top-up link, and the MCP tool result also carries the same link; each key has a per-task and per-day spending cap.
For service-oriented scenarios, a notify_url (must be https) can be passed when POSTing to /v1/jobs, and a single JSON POST callback is received after the task completes; also, sending an Idempotency-Key request header avoids accidentally starting a second task on retry.
All Facts (7)
Official measurement: calling render_preview for the example scenario with four lighting presets, under the default 5-second timeout it threw McpError: Timed out after 5.1 seconds; after changing to a 120-second timeout the same task returned a result in 7.3 seconds, actually consuming 4.03 GPU seconds.
50%UnverifiedJANCTION Render provides a remote MCP Server endpoint https://render.janction.jp/mcp, authenticated with a Bearer Token, obtained by calling curl -s -X POST https://render.janction.jp/v1/keys, with no registration required.
50%UnverifiedThrough the JANCTION Render MCP Server, the Agent can see 12 tools, with a typical call chain of scene_info → render_preview → render_estimate → render_final → render_status → render_download.
50%UnverifiedJANCTION Render gives each new key 500 yen of GPU time (5000 GPU seconds), valid for 14 days, with no card binding required; afterwards preview tasks get 2 GPU minutes free per day, and other GPU time is billed at 0.1 yen per GPU second.
50%UnverifiedOfficial measurement: a 24-frame 720p MP4 at 64 samples consumes about 58 GPU seconds, equivalent to about 6 yen.
50%UnverifiedWhen the JANCTION Render quota is exhausted, the task is rejected before running, the HTTP API returns 402 with a top-up link, and the MCP tool result also carries the same link; each key has a per-task and per-day spending cap.
50%UnverifiedFor service-oriented scenarios, a notify_url (must be https) can be passed when POSTing to /v1/jobs, and a single JSON POST callback is received after the task completes; also, sending an Idempotency-Key request header avoids accidentally starting a second task on retry.
50%