💻

Blender MCP

Updated June 2026⭐ Featured

The Blender MCP, built by ahujasid, provides blenderMCP connects Blender to an AI assistant over the Model Context Protocol, and it is two pieces rather than one: addon.py runs inside Blender and opens a socket server on port 9876 (BLENDER_HOST / BLENDER_PORT override it, so Blender can live on another machine), while the blender-mcp Python package is the MCP server your client launches and which relays commands into that socket. It is community-built and best for Coding & Dev.

by ahujasid

About

BlenderMCP connects Blender to an AI assistant over the Model Context Protocol, and it is two pieces rather than one: addon.py runs inside Blender and opens a socket server on port 9876 (BLENDER_HOST / BLENDER_PORT override it, so Blender can live on another machine), while the blender-mcp Python package is the MCP server your client launches and which relays commands into that socket. Blender must therefore be open with the 'Interface: Blender MCP' addon enabled and 'Connect to Claude' pressed in the BlenderMCP sidebar tab before any tool call can do anything — that missing step accounts for most 'it connects but nothing happens' reports. Requirements are Blender 3.0+ and Python 3.10+, installed via uv's official installer (not `pip install uv`, which may not create the uvx command clients need). Capabilities: read the scene graph and take viewport screenshots, create/modify/delete objects, apply materials and colours, run arbitrary Python through execute_blender_code, download models, textures and HDRIs from Poly Haven, search and pull Sketchfab models, and generate models with Hyper3D Rodin or Hunyuan3D — asset keys live in the addon preferences or in BLENDERMCP_* environment variables. Note two current caveats: the upstream GitHub repository does not resolve (github.com/ahujasid/blender-mcp redirects to github.com/MCPBlender/blender-mcp, which returned 404 on 2026-08-10) although the PyPI package is still being released, and anonymous telemetry is on by default and disabled with DISABLE_TELEMETRY=true. It is a third-party integration by Siddharth Ahuja, not made by Blender.

Unverifiable
not scored

No repository for this entry resolves against the live GitHub API, so there is nothing to verify. It is deliberately left unscored rather than given a number we cannot stand behind.

What could not be measured

These contributed nothing to the score — not a penalty, not a zero. They are why the confidence reads the way it does.

  • Source verificationunknown

    No repository for this entry resolves against the live GitHub API. Without a confirmed source there is nothing to inspect, so this entry is excluded from trust scoring entirely rather than scored on catalog metadata alone.

Installation

pip
uvx blender-mcp

blender-mcp confirmed live on PyPI — checked August 17, 2026.

BlenderMCP is two pieces, and people who install one and not the other account for most of the "it connects but nothing happens" reports. `addon.py` runs inside Blender and opens a socket server on port 9876; the `blender-mcp` Python package is the MCP server your client speaks to, which relays commands into that socket. So Blender has to be open, the addon has to be enabled, and you have to have pressed "Connect to Claude" in the BlenderMCP sidebar tab before any tool call can work. Requirements are Blender 3.0+ and Python 3.10+, and the install is via uv — the project is emphatic that you install uv with its official installer, not `pip install uv`, which may not create the `uvx` command your client needs.

Setting up BlenderMCP

  1. 1.Install uv properly

    macOS: `brew install uv`. Windows: `powershell -c "irm https://astral.sh/uv/install.ps1 | iex"`, then add %USERPROFILE%\.local\bin to your user PATH and restart the client. Linux: `curl -LsSf https://astral.sh/uv/install.sh | sh`, then open a new shell. Do not proceed before uv works.

    shell
    brew install uv   # macOS
  2. 2.Install and enable the Blender addon

    Download addon.py from the project, then in Blender: Edit → Preferences → Add-ons → Install…, select addon.py, and tick the box next to "Interface: Blender MCP". Note that the upstream GitHub repository currently does not resolve — github.com/ahujasid/blender-mcp redirects to github.com/MCPBlender/blender-mcp, which returns 404 as of this verification date — so blendermcp.org is the place to start looking for the addon file. The PyPI release is unaffected and current.

  3. 3.Add the server to your client

    Claude Desktop: Settings → Developer → Edit Config. Claude Code: `claude mcp add blender uvx blender-mcp`. Cursor takes the same JSON globally or in .cursor/mcp.json. On Windows, wrap it: `"command": "cmd", "args": ["/c", "uvx", "blender-mcp"]`. Run only one instance of the MCP server — Cursor or Claude Desktop, not both.

    claude_desktop_config.json
    {
      "mcpServers": {
        "blender": {
          "command": "uvx",
          "args": ["blender-mcp"]
        }
      }
    }
  4. 4.If the client cannot find uvx

    A GUI-launched client does not inherit your terminal PATH, so a bare `"command": "uvx"` can fail with `spawn uvx ENOENT` even though uvx works in a shell. Use the absolute path from `which uvx` / `where uvx` (e.g. /opt/homebrew/bin/uvx), then fully quit and relaunch the client — Cmd-Q on macOS, quit from the system tray on Windows.

  5. 5.Pin Python if you have conda, pyenv or asdf

    uv picks the interpreter, and on machines with an auto-activated conda base it can pick one that fails to build dependencies. Pin 3.11 and prefer uv-managed interpreters. If a failed attempt keeps replaying after you fix it, clear the cache: `uv cache clean blender-mcp && uvx --refresh blender-mcp`.

    mcp config
    {
      "mcpServers": {
        "blender": {
          "command": "uvx",
          "args": ["--python", "3.11", "blender-mcp"],
          "env": { "UV_PYTHON_PREFERENCE": "only-managed" }
        }
      }
    }
  6. 6.Start the connection inside Blender

    In the 3D View sidebar (press N), open the BlenderMCP tab, optionally tick Poly Haven to allow asset downloads, and click "Connect to Claude". Do not run the uvx command yourself in a terminal — the client starts it. Sometimes the first command does not go through and it works from the second onward.

What it can do, and the asset services behind it

Communication is a JSON protocol over TCP — commands carry a type and optional params, responses carry a status plus result or message. BLENDER_HOST (default localhost) and BLENDER_PORT (default 9876) let the MCP server talk to Blender on another machine.

Scene and object inspection

Read the scene graph and object details, including viewport screenshots so the model can see what it just built.

Object creation and modification

Create, modify and delete 3D objects; position, scale and orient them.

Material control

Apply and modify materials and colours on existing objects.

execute_blender_code

Runs arbitrary Python in Blender. This is the escape hatch that makes everything else possible and the single most dangerous tool here — always save your work first.

Poly Haven assets

Search and download models, textures and HDRIs through the Poly Haven API. Off unless you tick the checkbox in the addon panel, because it downloads files to your machine.

Hyper3D Rodin generation

Generate 3D models from a prompt. The free trial key is limited per day; beyond that, bring your own from hyper3d.ai or fal.ai.

Sketchfab search and download

Search and pull Sketchfab models, with your own Sketchfab API key.

Hunyuan3D

Additional generation backend, configured with a SecretId/SecretKey pair and an API URL.

What people use it for

Block out a scene from a description

Create a low poly scene in a dungeon, with a dragon guarding a pot of gold.

The canonical demo, and a fair test of whether the addon connection is live: you should see geometry appear in the viewport as tools run.

Dress a scene with real assets

Create a beach vibe using HDRIs, textures and models like rocks and vegetation from Poly Haven.

Requires the Poly Haven checkbox to be on. The project notes the model can be erratic about Poly Haven, so expect to steer it.

Fix look and framing

Make this car red and metallic, make the lighting like a studio, then point the camera at the scene and make it isometric.

Material and camera work is where the natural-language interface beats hunting through panels, and it is safe — no code execution needed.

Which one should you use?

Only run one MCP server against a single Blender instance — the addon exposes one socket.

Running it from Cursor and Claude Desktop at once

Do not. The project warns explicitly to run only one instance of the MCP server, in one client.

Hyper3D Rodin vs. Poly Haven

Poly Haven when a real asset exists — it is a library of CC0 models, textures and HDRIs. Hyper3D when nothing exists and you need something generated, accepting trial-key daily limits.

Driving a game engine instead

If the destination is a game engine rather than a render, an engine-side MCP server skips the export step entirely. Blender is the better answer for modelling, materials and stills.

Every command, environment variable, and endpoint above was read from the project’s own documentation on 2026-08-10: blender-mcp on PyPI (README, v1.8.0), BlenderMCP official site, Live check — GitHub repo redirect and PyPI release metadata.

Browse all MCP server setup guides.

Frequently Asked Questions

Why does Blender MCP connect but nothing happens in Blender?
The addon side is not running. Blender must be open with the "Interface: Blender MCP" addon enabled, and you must have clicked "Connect to Claude" in the BlenderMCP sidebar tab (press N to show the sidebar). Also do not run `uvx blender-mcp` in a terminal yourself — the client launches it. If the first command still fails, try again; the project documents that the first one sometimes does not go through.
Where do I download addon.py now that the GitHub repo 404s?
As of 2026-08-10, github.com/ahujasid/blender-mcp redirects to github.com/MCPBlender/blender-mcp and that URL returns 404 — the repository has been moved or taken private. The PyPI package is still being released (1.8.0, uploaded 2026-08-03), so `uvx blender-mcp` continues to work; blendermcp.org is the project's own site and the place to look for the current addon file. Treat any third-party mirror of addon.py with suspicion: it runs arbitrary Python inside Blender.
Why does uvx fail with spawn uvx ENOENT?
Your client was launched from a GUI and does not have uv on its PATH. Replace "uvx" with the absolute path from `which uvx` or `where uvx`, or on Windows use `"command": "cmd", "args": ["/c", "uvx", "blender-mcp"]`. Then fully quit and relaunch the client.
Can I run BlenderMCP with Blender on another machine?
Yes. Set BLENDER_HOST and BLENDER_PORT — for example BLENDER_HOST=host.docker.internal, BLENDER_PORT=9876. Default is localhost:9876.
Is execute_blender_code safe?
It runs arbitrary Python inside Blender, which is exactly as powerful and as dangerous as that sounds. The project says to use it with caution outside experiments and to always save your work first. Complex requests are also better broken into steps — timeouts are a documented failure mode for large single operations.
Does BlenderMCP send telemetry?
Yes, anonymous usage data by default. With consent checked it can include anonymised prompts, code snippets and screenshots; unchecked it collects only tool names, success/failure and duration. Turn it off entirely with DISABLE_TELEMETRY=true, either in the env block of your MCP config or on the command line.
Do I need API keys?
Not for core modelling. Asset and generation services need their own: store the Sketchfab and Hyper3D keys and the Hunyuan3D SecretId/SecretKey in Edit → Preferences → Add-ons → Blender MCP so they survive restarts, or inject them for headless use as BLENDERMCP_SKETCHFAB_API_KEY, BLENDERMCP_HYPER3D_API_KEY, BLENDERMCP_HUNYUAN3D_SECRET_ID, BLENDERMCP_HUNYUAN3D_SECRET_KEY and BLENDERMCP_HUNYUAN3D_API_URL.
Is BlenderMCP made by Blender?
No. The project states plainly that it is a third-party integration, not made by Blender. It is by Siddharth Ahuja.
What is Blender MCP?
Blender MCP is an MCP server built by ahujasid. BlenderMCP connects Blender to an AI assistant over the Model Context Protocol, and it is two pieces rather than one: addon.py runs inside Blender and opens a socket server on port 9876 (BLENDER_HOST / BLENDER_PORT override it, so Blender can live on another machine), while the blender-mcp Python package is the MCP server your client launches and which relays commands into that socket. Blender must therefore be open with the 'Interface: Blender MCP' addon enabled and 'Connect to Claude' pressed in the BlenderMCP sidebar tab before any tool call can do anything — that missing step accounts for most 'it connects but nothing happens' reports. Requirements are Blender 3.0+ and Python 3.10+, installed via uv's official installer (not `pip install uv`, which may not create the uvx command clients need). Capabilities: read the scene graph and take viewport screenshots, create/modify/delete objects, apply materials and colours, run arbitrary Python through execute_blender_code, download models, textures and HDRIs from Poly Haven, search and pull Sketchfab models, and generate models with Hyper3D Rodin or Hunyuan3D — asset keys live in the addon preferences or in BLENDERMCP_* environment variables. Note two current caveats: the upstream GitHub repository does not resolve (github.com/ahujasid/blender-mcp redirects to github.com/MCPBlender/blender-mcp, which returned 404 on 2026-08-10) although the PyPI package is still being released, and anonymous telemetry is on by default and disabled with DISABLE_TELEMETRY=true. It is a third-party integration by Siddharth Ahuja, not made by Blender.
Who built Blender MCP?
Blender MCP was built by ahujasid.
Is Blender MCP free?
Yes, Blender MCP has a free option. This MCP server is free and open-source. Check the GitHub repository for details.
How do I install Blender MCP?
Install Blender MCP with pip: uvx blender-mcp
What does Blender MCP integrate with?
Blender MCP integrates with Claude Desktop.
No verified public repositoryVisit Website

Repo Health

Actively maintained

Local/stdio install — runs on your machine, so there is no remote endpoint to verify live. Trust signal below is from the source repo.

Last commit
1mo ago
Install
pip
Package
blender-mcp

Quick Info

Install Type
pip
Author
ahujasid
Categories
2
Integrations
1

Related Servers

💻

Everything

Reference/test server with prompts, resources, and tools. Perfect for testing MCP implementations.

Local
💻

Git

Tools to read, search, and manipulate Git repositories. Full Git operations support.

Local
🤖

Sequential Thinking MCP Server

a single structured-reasoning tool that lets a model plan, revise and branch its own chain of thought instead of answering in one shot. Published by Anthropic as part of the official modelcontextprotocol/servers monorepo (89,000+ stars, actively maintained), it exposes exactly one tool — sequential_thinking — and that tool is the whole product. Each call carries a `thought` string plus bookkeeping fields: `thoughtNumber`, `totalThoughts`, and `nextThoughtNeeded`, which the model flips to false when it is done. The interesting fields are the optional ones. `isRevision` and `revisesThought` let the model go back and correct an earlier step rather than plowing ahead on a bad assumption; `branchFromThought` and `branchId` let it fork into an alternative line of reasoning and carry both forward; `needsMoreThoughts` lets it extend past its own original estimate when a problem turns out to be deeper than it looked. In practice you never call the tool by hand. You connect the server to an MCP host and ask a question that deserves more than one pass — plan a PostgreSQL 14 to 16 migration and revise if downtime exceeds five minutes, work out why a deploy only fails in production, compare three architectures and branch when an assumption breaks. You can tell it is working when the host inspector shows repeated sequential_thinking calls with a rising `thoughtNumber` rather than a single response. Install with `npx -y @modelcontextprotocol/server-sequential-thinking` — note the hyphenated package name, which differs from both the `sequentialthinking` directory in the repo and the Docker image `mcp/sequentialthinking`, a mismatch that breaks a lot of copied configs. A Docker image is published alongside the npm package, and the README carries one-click VS Code install buttons for both transports. Set `DISABLE_THOUGHT_LOGGING=true` if you do not want every thought written to the server log.

Local
💻

21st.dev Magic

Create crafted UI components inspired by the best 21st.dev design engineers.

Local
💻

GitHub MCP Server

authenticated access to the whole GitHub platform — repositories, files, branches, issues, pull requests, Actions runs, security alerts, discussions and notifications — from Claude, Cursor, VS Code, Copilot CLI and any other MCP host. There is no npm package for this server, and that trips up most people who try to install it: `@github/mcp-server` is not published to the npm registry, so any `npx` line you find for it will fail. GitHub ships it three other ways. The easiest is the hosted remote server at https://api.githubcopilot.com/mcp/, which needs no install at all — point an HTTP-transport MCP client at that URL and log in with OAuth (VS Code 1.101+, Claude Desktop, Claude Code, Cursor and Windsurf all support this). The second is the official Docker image ghcr.io/github/github-mcp-server, which is what the copy-paste command on this page runs; on github.com it now performs a browser-based OAuth login on first use and keeps the token in memory only, which is why the published Docker configs map a fixed loopback callback port (-p 127.0.0.1:8085:8085 with GITHUB_OAUTH_CALLBACK_PORT=8085) so the container can receive the callback. Prefer a token? Set GITHUB_PERSONAL_ACCESS_TOKEN instead — it takes precedence over OAuth, and the minimum useful scopes are repo, read:org and read:packages. The third is the native Go binary from the repository's releases, which needs no fixed port for the OAuth flow. GitHub Enterprise Server has no hosted option: use the local server with --gh-host or GITHUB_HOST set to your instance (include the https:// scheme — it defaults to http://, which GHES rejects). Toolsets can be narrowed with GITHUB_TOOLSETS, and an insiders channel is available at /mcp/insiders or via the X-MCP-Insiders header.

Auth required📘

Sponsored

Better Stack

Free Plan

Get alerted when your APIs, browser tests, payment pipelines, or MCP server dependencies go down. Used by 100K+ developers.

Start monitoring free →