Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Natron MCP — AI Agent Control of Natron Compositor

    1

    Verified on Windows (portable build) and Linux (AppImage, WSL2/Xvfb).

    $15

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Natron MCP — AI Agent Control of Natron Compositor

    Natron MCP — AI Agent Control of Natron Compositor

    Example session with this skill installed

    Connect to my headless Natron, load rain_shot.%04d.exr, add a grade node lifting the mids by 0.15, and render to rain_shot_v2.mov

    • Read your context and instructions
    • Compiled the natron mcp

    Agent installs natron-mcp, starts NatronRenderer with the bridge, connects over MCP, builds Read (EXR sequence) → Grade (master lift 0.15) → Write (MOV), calls render_start, polls render_status until frames finish, and reports the output path and frame count.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    natron-mcp is a two-part Python system giving AI agents full, safe access to the Natron compositor (the Natron 3D 2.6+ fork by Niik-l). A lightweight TCP bridge runs inside Natron (GUI or headless NatronRenderer) and exposes an NDJSON op protocol; a FastMCP server discovers running sessions and forwards 26+ agent-facing tools: project create/save/load, node create/connect/delete, readers/writers, knobs and parameters, expressions, keyframes, render start/status, plugin search, session spawn/kill/restart, modal detection, and an AST-gated exec_python (deletion, subprocess, eval, file writes and network-without-timeout are blocked). Every op is audit-logged.

    This skill walks your AI agent through the whole lifecycle: verifying the environment, downloading the right release for the platform, installing the bridge into a portable Natron install or a fork build tree, starting a headless or GUI session, connecting an MCP client, building and rendering a first node graph, and troubleshooting every known failure mode.

    Verified on Windows (portable build) and Linux (AppImage, WSL2/Xvfb). GPL-2.0, matching Natron's own license.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Workflow Automation.

    Targets the Natron 3D 2.6+ fork (Niik-l); upstream Natron is untested. OS-level modal dismissal is Windows-only (Linux degrades gracefully to the in-host watchdog). Pixel-preview inline images are limited; data ops return file paths. No timeline-animation authoring beyond keyframe set/remove ops. No macOS testing.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 4 days ago

    Accounts
    GitHub0
    Software
    Natron 3D 2.6+ (Niik-l fork)Python 3.10+MCP-capable AI clientNatron MCP
    Optional
    Git (to clone the repo)xvfb (Linux GUI tests)

    The bridge runs on Natron's bundled Python; the MCP server needs its own Python 3.10+. On Linux, extract the AppImage once with --appimage-extract and install the listed distro packages (see skill).

    Permissions

    Allowed Hosts

    github.com
    127.0.0.1

    File Scopes

    ~/.Natron/natron-mcp/
    Natron
    install
    dir

    The agent downloads the natron-mcp repo from GitHub, runs its installer against the local Natron install, and talks to the bridge/server on 127.0.0.1 only. Sessions are audit-logged under ~/.Natron/natron-mcp/.

    Listed4 days ago

    What's inside

    Frequently Asked Questions

    It is built and verified against the Natron 3D 2.6+ fork (Niik-l). Upstream releases may work with minimal effort but are untested.

    Yes: a Natron 3D 2.6+ (fork) install and Python 3.10+. The skill itself is free; the natron-mcp source it installs is GPL-2.0 from GitHub.

    exec_python is AST-gated on both server and bridge: deletion, subprocess, eval/exec, file writes, and network calls without timeout are blocked; every op is audit-logged to ~/.Natron/natron-mcp/audit-<session>.jsonl. Set NATRON_MCP_DISABLE_EXEC=1 or NATRON_MCP_READONLY=1 for stricter modes.

    Windows (portable build) fully verified; Linux AppImage verified 35/35 under WSL2/Xvfb. No Node.js runtime needed.

    Both. The same MCP server and tools work unchanged against GUI Natron and headless NatronRenderer sessions; sessions_spawn can start hosts for you.

    session_modal_status detects the dialog (even during a full wedge) and session_modal_dismiss clears it; a watchdog auto-answers per NATRON_MCP_MODAL_POLICY; session_restart is the last resort.

    python install.py --natron <root> --action remove --yes — removal is manifest-driven and exact, leaving the rest of the Natron install untouched.

    GPL-2.0, matching Natron's own license (the bridge loads inside Natron).

    New in Workflow Automation