Claude for Creative Work: MCP Connectors for Blender, Adobe, Ableton, and More

✍️ OpenClawRadar📅 Published: April 28, 2026🔗 Source
Claude for Creative Work: MCP Connectors for Blender, Adobe, Ableton, and More
Ad

Anthropic has announced Claude for Creative Work, a set of MCP (Model Context Protocol) connectors that let Claude interface directly with professional creative software. The connectors are released in partnership with Blender, Autodesk, Adobe, Ableton, and Splice, among others.

Specific Connectors

  • Ableton: Answers grounded in official Live and Push documentation.
  • Adobe for creativity: Enables image, video, and design generation across 50+ Creative Cloud apps (Photoshop, Premiere, Express, etc.).
  • Affinity by Canva: Automates batch image adjustments, layer renaming, file export, and generates custom features in-app.
  • Autodesk Fusion: Create and modify 3D models through conversational Claude (subscription required).
  • Blender: Natural-language interface to Blender's Python API. Lets Claude explore complex setups, access docs, and add new tools to Blender's interface. Anthropic joined the Blender Development Fund as a patron.
  • Resolume Arena/Wire: VJs and live visual artists can control Arena, Avenue, and Wire via natural language for live performance.
  • SketchUp: Describe a room or concept, then open it in SketchUp for refinement.
  • Splice: Search the royalty-free sample catalog from within Claude.
Ad

Use Cases

  • Learning tools: Ask Claude to explain modifier stacks, synthesis techniques, or unfamiliar features.
  • Extending tools with code: Claude Code writes scripts, plugins, and generative systems (shaders, procedural animations, parametric models).
  • Bridging tools: Translate formats, restructure data, sync assets across applications without manual handoffs.
  • Rapid exploration: Claude Design (Labs product) visualizes UI/UX options and exports to Canva.
  • Repetitive production: Batch-process assets, set up project scaffolding, apply procedural changes across scenes.

Blender-Specific Details

The Blender MCP connector allows analyzing and debugging entire scenes, building custom scripts with Blender's Python API, and batch-applying changes. Because it's built on MCP, the connector is accessible to other LLMs besides Claude.

Educational Partnerships

Anthropic is working with three programs: Art and Computation at RISD, Fundamentals of AI for Creatives at Ringling College of Art and Design, and the MA/MFA Computational Arts at Goldsmiths, University of London.

📖 Read the full source: HN AI Agents

Ad

👀 See Also

Quanta-SDK v0.9.2 adds MCP server for quantum circuit execution via AI agents
Tools

Quanta-SDK v0.9.2 adds MCP server for quantum circuit execution via AI agents

Quanta-SDK v0.9.2 now includes an MCP (Model Context Protocol) server that provides AI agents like Claude or GPT with tools to execute and interpret quantum circuits. The server offers over 20 tools including circuit execution on IBM hardware, result interpretation, noise analysis, and quantum financial pricing.

OpenClawRadar
OpenClaw Context Meter Plugin Shows Telegram Token Usage Percentage
Tools

OpenClaw Context Meter Plugin Shows Telegram Token Usage Percentage

A new OpenClaw plugin displays token usage percentage after every Telegram bot response, showing values like '45k / 200k (22%)' and detecting compaction events. The plugin avoids OOM issues by hardcoding context windows instead of using execSync.

OpenClawRadar
Merlin: Local-first LLM context dedup – measure up to 71% chunk overlap, free & open-core
Tools

Merlin: Local-first LLM context dedup – measure up to 71% chunk overlap, free & open-core

Merlin is a local-first context dedup tool that measured 22-71% chunk overlap across 22M passages from real agent/RAG sessions. Ships as HTTP proxy (Ollama/vLLM/SGLang/llama.cpp), MCP server (Claude/Cursor/OpenClaw), or standalone CLI. MIT open-core with daily usage caps.

OpenClawRadar
Code retrieval for AI agents: Why vector embeddings fail and per-file LLM graphs win
Tools

Code retrieval for AI agents: Why vector embeddings fail and per-file LLM graphs win

After a year of building a code indexing system, the team behind Bytebell found that vector embeddings on code chunks and Tree-sitter ASTs both fell short, while per-file LLM summaries stored in a Neo4j graph with semantic fulltext search significantly improved retrieval precision.

OpenClawRadar