Use it from Python
Build render pipelines, tests, and batch jobs around import bpy in a normal Python workflow.
Reference image → agent → headless Blender
Watch an agent rebuild images without opening the GUI.
It reads the reference, writes Python through the CLI, renders headlessly, inspects the result, and iterates.
Open community preview
Give your agent a reference image. Let it build the Blender scene.
blender-mcp makes Blender a headless backend for coding agents. The agent uses the image as its goal, drives Blender through a narrow CLI and MCP surface, then renders and inspects its own work as it goes.
Standalone bpy packaging remains work in progress. Track build work in PR #9.
01 / What developers can do
Blender is the backend. Your agent does the work.
Use the same Python runtime directly, or expose a small set of operations to an agent.
Let an agent drive it
Give an agent four CLI-friendly MCP tools to write Python, render the scene, inspect the image, and keep working.
02 / How it works
Reference → Build → Inspect
The loop is visual, but the backend is headless. No one has to sit in front of Blender.
- 01
Reference
The agent gets an image and a goal—rebuild what it sees as a real Blender scene.
- 02
Build
It writes Python through the CLI and runs Blender as a headless backend.
- 03
Inspect
It renders its own work, looks at the result, and keeps going. The GUI never needs to open.
03 / Initial evidence
Measured in renders, not claims.
One transparent community-preview run on the public BlenderBench dataset.
- 27
- public BlenderBench tasks
- 270
- GPU-rendered and GPU-CLIP-scored checkpoints
- 4
- exactly four MCP tools
10 rounds per taskno VLM judgeno score feedback during generation
Video description and results
Text alternative for the silent compilation
The compilation covers 27 tasks and 270 rounds. Each checkpoint identifies its task and round, places the generated scene beside its target render, and reports the CLIP image-embedding cosine similarity and generated Python token count shown on screen.
Category results
Mean best distance by difficulty
Category mean best N-CLIP ×100. N-CLIP is derived from CLIP image-embedding cosine similarity; lower is better. Values retain the precision of the recorded aggregate.
Category mean best N-CLIP times 100
| Category | N-CLIP ×100 |
|---|---|
| Camera / Level 1 | 0.3738476170433892 |
| Level 2 | 1.603147718641493 |
| Level 3 | 2.2820386621687145 |
203e4d325e9438ca55b29bdfc4f6a90842d74e68, licensed CC BY 4.0.Open community preview