tfmstyle-blender-tools
You can inspect and run TFMSTYLE Blender addons in a running Blender through three MCP tools.
List TFMSTYLE addons currently running in Blender and their versions.
Describe an addon's actions/settings, current values, allowed ranges, per-item
listssettings, and guide.Apply settings, run an addon action, or both; settings persist like moving a slider.
Validate all inputs first: one bad value rejects the whole call and changes nothing.
Require
confirm=truefor actions that clear/reset things, and refuse starting an action already running.Measure result changes (size, vertices, lowest point, keyframes) and warn on huge meshes or no-op runs.
Only run addon-defined actions; import/export/file/online actions are excluded.
Not for general Blender operations like deleting/moving objects, setting keyframes, or changing frames; use Blender's own MCP tools for those.
Provides tools for interacting with a running Blender instance to list, describe, and run enabled TFMSTYLE addons such as Rock, Orbiter, and Graph. It can read addon actions, settings, current values, and allowed ranges live from Blender, then apply settings and run safe addon actions with validation and confirmation for destructive/reset operations.
Click on "Deploy 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., "@tfmstyle-blender-tools list the TFMSTYLE Blender addons currently enabled and their versions"
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.
Blender Agent Tools
One MCP toolbox that lets any agent use the TFMSTYLE Blender addons inside a running Blender.
This is a prototype. It covers Rock, Orbiter and Graph. The other addons join as they pass the same test.
The three tools
Tool | What it does |
| The TFMSTYLE addons running in Blender, with their versions |
| One addon's actions and settings, with current values and allowed ranges |
| Applies settings, runs an action, or both |
Settings and ranges are read from Blender at the moment of the call. Nothing is installed in Blender, and the addons themselves do not change.
describe_addon leaves out the addon's private state, marks which settings apply to the current choices, and adds a short guide. The guides in guides.py are the one part kept by hand, so check them when an addon changes.
run_addon checks everything first, and one bad value refuses the whole call. It only runs an addon's own actions. It leaves out anything that imports, exports, opens a file or goes online, an action that clears or resets something needs confirm=true, and starting something already running is refused.
ok only says the action ran. The reply also measures what changed: size, vertices, lowest point and keyframes, with a warning for a huge mesh or a run that changed nothing.
These tools only run the addons. Use Blender's own MCP tools to delete or move objects, set keyframes, or change the frame.
Related MCP server: Blender Open MCP
What you need
Blender 5.1 or newer, with the TFMSTYLE addons you want enabled, separately or through the Addon Collection
The MCP add-on from Blender Lab, enabled with its server started
Python 3.10 or newer
Blender Lab warns that its server runs agent written code with no guards. The checks here cover these three tools only. They do not limit what an agent can do through Blender Lab's own tools.
Install
pip install git+https://github.com/TFMSTYLE/Blender_Agent_Tools.gitThen add it to your MCP client. For most clients:
{
"mcpServers": {
"tfmstyle-blender-tools": {
"command": "tfmstyle-blender-agent-tools"
}
}
}It uses the same BLENDER_MCP_HOST and BLENDER_MCP_PORT settings as Blender Lab's server, and both can run at once.
With no window
Background Blender works too, with blender --background --online-mode --command blender_mcp. Actions that keep running, like Orbiter's orbit, are refused there because they need an open window.
Test
python tests/live_test.py <folder with the three zips> <collection zip>It starts throwaway Blenders with their own settings folder and port, and drives them through a real MCP connection. A Blender you have open is left alone.
License
GPL 3.0 or later.
Available Tools
3 toolsdescribe_addonBRead-only
Show one addon's actions and settings, with current values and allowed ranges.
Settings kept per item appear under lists, for example one set per saved graph.
Set those with the name list[index].setting.
Inputs: addon (required).
list_addonsARead-only
List the TFMSTYLE addons running in Blender, with their versions.
Inputs: no named parameters.
run_addonADestructive
Apply settings to an addon, run one of its actions, or both.
Everything is checked first: one bad value refuses the whole call and changes nothing. An action that clears or resets something needs confirm=true. Settings stay as set, the same as if a person had moved the slider.
Inputs: addon (required), action, confirm, settings.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 3 tools
Each tool targets a distinct operation: list_addons enumerates addons, describe_addon inspects one addon's actions/settings, and run_addon applies settings or executes actions. There is no meaningful overlap in purpose, and the descriptions make the boundaries clear.
All three tools follow a consistent verb_noun pattern: list_addons, describe_addon, run_addon. The naming is predictable and easy to map to the intended operations.
Three tools are well-scoped for a server focused on managing TFMSTYLE Blender addons: one to enumerate, one to inspect, and one to act. Each tool earns its place without redundancy.
The surface covers the core lifecycle of discovering addons, inspecting their settings/actions, and running them with validation and confirmation. Minor gaps exist, such as enabling/disabling addons or viewing logs, but core workflows are covered.
Related MCP Connectors
Blender-as-a-service for agents: search 3D assets, run Blender Python, or brief the studio agent.
Connect your AI assistant to Blender for 3D modeling, Geometry Nodes, materials, lighting, animation, rendering and export. Inspect real scenes and viewport images, refine the result and restore checkpoints. OAuth connects your chat to Blender Bridge Desktop on an online Windows 10/11 x64 computer running Blender 4.2+. The chat can run on another device. Requires an active Blender Bridge account; 7-day trial available. AI model access is separate.
Discover, inspect and run 63,000+ agent tools from one balance. Pay per call, no subscriptions.
A registry of AI agent tools — MCP servers, APIs, CLIs, SDKs — kept current by automated ingestion.
Related MCP Servers
- AlicenseAqualityBmaintenanceBuild Blender scenes through 4 tools and 338 tokens of fixed context: on-demand operation schemas, batches of up to 100 edits, preview captures and permission-gated Python for full bpy access.418MIT
- FlicenseNot gradedqualityAmaintenanceEnables AI agents to control a live Blender session, inspect and modify scenes, manage PolyHaven assets, render images, and route prompts through configurable local or OpenAI-compatible LLM backends.-
- FlicenseNot gradedqualityAmaintenanceEnables AI agents to control a live Blender session, including scene and object operations, materials, rendering, PolyHaven asset downloads, and prompts to configurable LLM backends through MCP tools.-
- AlicenseBqualityBmaintenanceEnables AI clients to control Blender through its official MCP add-on: running arbitrary bpy code or background CLI jobs, auditing scenes and blend files, searching Blender API and manual docs, capturing screenshots and renders, and searching or importing free Poly Haven models, HDRIs and textures at real-world scale. All 28 tools are served over one stdio connection to a single Blender instance.28GPL 3.0