Provides headless Blender automation for inspecting .blend scenes and rendering preview stills without modifying source scenes.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Blender MCP inspect the scene in scene.blend"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A headless-first Model Context Protocol server for safe, deterministic Blender automation.
This repository is under active development. The first milestone is deliberately small:
discover and verify Blender;
inspect a .blend scene through a versioned JSON contract;
render a preview still without mutating the source scene; and
expose those capabilities as typed MCP tools.
The project does not accept arbitrary Python or shell commands from MCP clients.
Related MCP server: Blender MCP Bridge
Headless Blender is the authoritative execution engine.
MCP tools are thin wrappers around testable application services.
Blender and the MCP server run in separate Python environments.
Every Blender request uses a versioned, allowlisted JSON envelope.
Files are restricted to one configured project root.
Source scenes are not overwritten by inspection or preview operations.
Failures return stable codes and request IDs.
macOS Apple Silicon
Blender 5.2 LTS
Python 3.12
MCP Python SDK 2.x
Additional platforms and an interactive Blender add-on are intentionally deferred until the headless path is reliable.
src/blender_mcp/ MCP server and application core
blender_adapter/ bundled scripts executed by Blender
schemas/ versioned JSON contracts
fixtures/ reproducible Blender fixtures
tests/ unit, security, integration, and end-to-end tests
docs/ contracts and architecture notescd /Users/akki/Desktop/github/blender-mcp
uv sync --python 3.12 --extra devChoose a project root containing the .blend files and preview outputs that
the server is allowed to access:
export BLENDER_MCP_PROJECT_ROOT=/absolute/path/to/blender-project
export BLENDER_MCP_BLENDER_EXECUTABLE=/Applications/Blender.app/Contents/MacOS/Blender
uv run blender-mcpThe project root is a security boundary. MCP paths must be relative to it; absolute paths, traversal, symlink escapes, and unexpected file extensions are rejected.
[mcp_servers.blender]
command = "/opt/homebrew/bin/uv"
args = [
"run",
"--project",
"/Users/akki/Desktop/github/blender-mcp",
"blender-mcp",
]
cwd = "/Users/akki/Desktop/github/blender-mcp"
startup_timeout_sec = 20
tool_timeout_sec = 300
default_tools_approval_mode = "writes"
required = false
[mcp_servers.blender.env]
BLENDER_MCP_PROJECT_ROOT = "/absolute/path/to/blender-project"
BLENDER_MCP_BLENDER_EXECUTABLE = "/Applications/Blender.app/Contents/MacOS/Blender"Restart Codex after changing MCP configuration.
get_blender_status
inspect_scene
render_preview
The server deliberately has no arbitrary Python, shell, add-on installation, URL download, or external-asset tool.
Run tests that do not require Blender:
uv run pytest tests/unit tests/security tests/contract -qRun the real Blender integration suite:
uv run pytest tests/integration/test_blender_adapter.py -qOn the reference macOS machine, Blender 5.2 can crash during native startup
when launched inside the Codex application sandbox. The identical test passes
when run through an approved unsandboxed Blender process or directly from
Terminal. The launcher classifies an early segmentation fault as
BLENDER_CRASHED; it is not reported as an adapter validation failure.
See the frozen M0 contract for the exact boundaries, schemas, errors, and deferred capabilities.
Check whether the configured Blender and bundled adapter are available.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true, idempotentHint=true, and destructiveHint=false. The description adds 'configured' and 'bundled' context, indicating the check covers configuration and adapter availability. However, it does not disclose additional behaviors such as output format or potential latency, though the output schema exists.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single concise sentence that front-loads the verb and resource. No unnecessary words or redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple no-parameter status check, the description is complete: annotations cover safety, output schema covers return values, and the description explains what is being checked. It could be slightly more explicit about what 'available' means, but overall it is sufficient given the tool's simplicity.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool has zero parameters, receiving the baseline score of 4. The description does not need to explain parameter semantics, and the schema coverage is complete by default.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the function with a specific verb ('check') and resource ('configured Blender and bundled adapter availability'). It distinguishes itself from siblings (inspect_scene, render_preview) by focusing on environment readiness rather than scene operations.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The usage context is implied as a prerequisite check before other Blender operations, but there is no explicit guidance on when to use it versus alternatives or any exclusions. The description does not reference sibling tools or provide when-to-use/when-not-to-use instructions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Inspect a project-relative .blend file without modifying it.
| Name | Required | Description | Default |
|---|---|---|---|
| scene_path | Yes |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true and destructiveHint=false, and the description adds 'without modifying it', reinforcing the safe read-only nature. It also adds the 'project-relative' scope, which informs the agent about path resolution behavior. No contradiction with annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, front-loaded sentence with no filler. Every word contributes to understanding the tool's purpose and scope.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple read-only inspection tool with one parameter and an output schema, the description covers the essential context: what the tool does and the relation of the path to the project. It doesn't mention any prerequisites or failure modes, but those are not critical given the tool's simplicity.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema has no descriptions for scene_path, and the description compensates by indicating the path is a project-relative .blend file. This gives the single parameter more meaning than the schema alone, though it could be more explicit about path format or file existence requirements.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's function with a specific verb ('Inspect') and resource ('a project-relative .blend file'), and the constraint 'without modifying it' distinguishes it from mutation tools. It also differentiates from sibling tools like render_preview by focusing on inspection rather than rendering.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage context (inspect a .blend file without modifying it) but provides no explicit when-to-use or alternatives. There is no mention of scenarios where another tool like render_preview or get_blender_status would be more appropriate, leaving the agent to infer from the tool's name.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Render a PNG preview to a project-relative output path.
| Name | Required | Description | Default |
|---|---|---|---|
| max_width | No | ||
| max_height | No | ||
| scene_path | Yes | ||
| output_path | Yes | ||
| allow_overwrite | No |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=false and destructiveHint=false, signaling a non-read-only but non-destructive operation. The description adds that it writes a PNG to a project-relative path, which clarifies the output side effect. However, it does not mention overwrite behavior (despite allow_overwrite parameter) or any dependencies on scene state, leaving room for more transparency.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, clear, front-loaded sentence with no wasted words. It is appropriately concise for the core purpose, though the lack of additional details makes it slightly under-specifying for the tool's complexity, so it does not earn a 5.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a tool with 5 parameters, 2 required, sparse annotations, and an output schema, the description is too minimal. It does not explain when to use it relative to siblings, prerequisites like an open scene, or the purpose of parameters. The agent would need to guess from the schema, which lacks descriptions.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 0%, so the description must compensate for the lack of parameter explanations. It only hints that output_path should be project-relative, but does not explain scene_path, max_width, max_height, or allow_overwrite. This is insufficient for a 5-parameter tool with no other parameter documentation.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description 'Render a PNG preview to a project-relative output path' clearly specifies the verb (render), the resource (PNG preview), and the destination (project-relative output path). This distinguishes it from siblings like get_blender_status and inspect_scene, which focus on status and inspection rather than rendering.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies this tool is for generating preview images, while sibling tools handle status checking and scene inspection, but it does not explicitly state when to use this tool over alternatives or any exclusions. The requirement that output path be project-relative is a constraint but not a usage guideline versus alternatives.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v0.1.0get_blender_statusinspect_scenerender_previewEach tool has a clearly distinct purpose: checking environment status, inspecting a blend file, and rendering a preview. No two tools overlap in function.
All tool names follow a consistent verb_noun pattern (get_blender_status, inspect_scene, render_preview), making the intended action obvious.
Three tools is at the low end of the typical well-scoped range, but each serves a distinct function without redundancy. The count feels minimal rather than excessive.
The server lacks any tools for creating, editing, or saving Blender scenes, which are common operations for a Blender integration. The existing tools only cover a narrow inspect-and-render workflow, leaving significant gaps for general-purpose use.
Cloud Blender for AI agents: scenes, assets, renders, MP4, STL, GLB — over hosted remote MCP.
A Model Context Protocol server for Wix AI tools
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
Model Context Protocol server for Studex tools, notifications, and profile integrations
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/YellowFoxH4XOR/blender-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server