Blender OpenBLOCK
MCP for Blender Using Open Weights Models
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Local-first Model Context Protocol (MCP) server for controlling a live Blender session from AI agents, with a provider-agnostic LLM backend:
- Ollama
- LM Studio
- llama.cpp server
- OpenAI (and every OpenAI-compatible endpoint: vLLM, TGI, OpenRouter,
Groq, Together, Azure AI Foundry via its OpenAI-compatible surface, ...)
- any server that speaks
POST /chat/completionswith an optional base URL
and API key
Providers are configured at startup (CLI flags or environment variables) and can be switched at runtime through MCP tools, so an agent can hop between backends without restarting the server.
MCP Client ──▶ FastMCP Server ──▶ (TCP 9876) ──▶ Blender add-on (addon.py) ──▶ bpy │ └──────────────▶ (HTTP) ──▶ LLM provider (Ollama / LM Studio / llama.cpp / OpenAI-compatible / Azure)
Repository layout
Quick start
1. Install
Requires Python ≥ 3.10.
python -m venv .venv .venv/Scripts/pip install -e ".[dev]" # Windows (bash/PowerShell) # or: source .venv/bin/activate && pip install -e ".[dev]" # macOS/Linux
2. Enable the Blender add-on
- Open Blender.
Edit → Preferences → Add-ons → Install..., chooseaddon.py.- Enable Blender MCP.
- In the 3D Viewport press
N, open the Blender MCP tab and click
**Start MCP
9984471ca7c1OBSERVED · 2026-10-07Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add blender-open-mcp --env BLENDER_OPEN_MCP_API_KEY=${BLENDER_OPEN_MCP_API_KEY} -- uvx blender-open-mcp{
"mcpServers": {
"blender-open-mcp": {
"command": "uvx",
"args": [
"blender-open-mcp"
],
"env": {
"BLENDER_OPEN_MCP_API_KEY": "${BLENDER_OPEN_MCP_API_KEY}"
}
}
}
}Exposed tools (19)
8 read · 10 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
blender_ai_prompt | read | |
blender_create_object | write | Create a new primitive mesh object in the Blender scene. |
blender_delete_object | destructive | Permanently delete an object from the current Blender scene. Destructive. |
blender_download_polyhaven_asset | write | Download a PolyHaven asset and import it into the active Blender scene. |
blender_execute_code | write | Execute arbitrary Python code inside Blender using the bpy module. Destructive. |
blender_get_llm_provider | read | Return the active LLM provider configuration (API key masked). |
blender_get_object_info | read | Retrieve detailed information about a specific Blender object by name. |
blender_get_ollama_models | read | List models available from the active Ollama endpoint (alias for blender_list_llm_models). |
blender_get_polyhaven_categories | read | Fetch the list of available asset categories from the PolyHaven API. |
blender_get_scene_info | read | Retrieve a full summary of the current Blender scene (objects, camera, render settings). |
blender_list_llm_models | read | |
blender_modify_object | write | Modify an existing Blender object |
blender_render_image | write | Trigger a render of the current Blender scene and save the result to disk. |
blender_search_polyhaven_assets | read | Search the PolyHaven asset library by type and optional category filters. |
blender_set_llm_provider | write | |
blender_set_material | write | Create a Principled BSDF material (reusing by name) and assign it to an object. |
blender_set_ollama_model | write | Set the Ollama model used by the LLM backend. |
blender_set_ollama_url | write | Update the Ollama server base URL used by the LLM backend. |
blender_set_texture | write | Apply a previously downloaded PolyHaven texture to a Blender object. |
Trust audit
BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (1 observation(s))
- Network
- declared (5 observation(s))
- Shell
- declared (1 observation(s))
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (2)
exec(compile(code, "<blender_mcp>", "exec"), local_ns)
blender_delete_object
Gates applied: no_behavioural_pass.
9984471ca7c1full audit observations/trust-audit/mcp-server/dhakalnirajan__blender-open.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-07 | 9984471ca7c1 | BLOCK | D | 69 | first audit |
Questions
What is the Blender Open MCP server?
MCP for Blender Using Open Weights Models
What tools does Blender Open expose?
19 in total: 8 read-only, 10 that write, and 1 that can delete or overwrite (blender_delete_object). Every one is listed on this page with its risk.
Is Blender Open safe to connect to an agent?
No — not without reading the findings first. The audit graded it D (69/100) and found 1 critical or high issue in the source. Each one is listed on this page with the file and line it is on. Separately from the audit: 1 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does Blender Open need?
It reads BLENDER_OPEN_MCP_API_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does Blender Open run?
It speaks stdio and streamable-http, so it runs as a local process your client starts. It is published on PyPI as blender-open-mcp.
How current is this page?
The grade is for one exact copy of the source (9984471ca7c1), read on 2026-10-07. The repository is watched and re-audited when it changes.