Official Cloudinary MCP server for uploading, transforming, analyzing, and organizing media assets in Cloudinary's image and video CDN.

MCP server for searching and retrieving high-quality stock photos from Unsplash with proper attribution metadata.

MCP server for YouTube Data API integration, enabling video search, transcript retrieval, and channel data access.

Official ElevenLabs MCP server. Generate natural-sounding speech, clone voices, and produce multilingual audio from AI agents using state-of-the-art TTS.

Run image generation, video synthesis, and computer vision models at scale via fal.ai. Fast async inference for Flux, SDXL, Stable Video Diffusion, and more.

Generate, edit, and upscale images using Stability AI models including Stable Diffusion 3 and Stable Image Ultra. Control generation with precise prompts and style parameters.

Real-time and batch speech-to-text transcription via Deepgram Nova-2. Transcribe audio files, detect speakers, and extract keywords from voice recordings in agent workflows.

Search and retrieve high-quality free stock photos and videos from Pexels. Access millions of curated images and clips for content pipelines, marketing, and media production.

Access Spotify's music catalog, playlists, listening history, and recommendations. Search tracks, albums, and artists. Control playback and manage user playlists via the Spotify Web API.

Enterprise-grade speech intelligence. Transcribe audio and video with speaker diarization, sentiment analysis, topic detection, PII redaction, and auto chapters. Best-in-class accuracy.

Transcribe audio to text using OpenAI Whisper. Supports 99 languages, translation to English, and timestamp generation. Run locally or via OpenAI API for production transcription.

Mux video infrastructure via MCP. Upload, transcode, and stream video at scale — plus analytics, thumbnails, and live streaming from agents.

Official Excalidraw MCP App. Stream hand-drawn diagrams with viewport control and interactive fullscreen editing inside MCP-compatible clients.

Publish and manage content on Ghost via MCP. Create posts, manage members, send newsletters, and read publication analytics — integrating Ghost's headless CMS into agent-driven content workflows.

Official Blender MCP server for AI-powered 3D modeling, scene analysis, and Python scripting within Blender.

Official Adobe Express developer MCP server for building add-ons and integrating with Adobe Creative Cloud workflows.

MCP server for controlling Ableton Live — create tracks, edit MIDI, manage sessions, and automate music production via AI.

MCP server for Resolume Arena and Avenue with 206 tools for playback, composition control, Advanced Output, and show recovery.

MCP server for SketchUp 3D modeling — create geometry, manage scenes, apply materials, and export models with AI assistance.

MCP server connecting AI agents to 621,500+ real app screens from Mobbin's design reference library. Search by app name, screen type, UI pattern, or flow to get real-world design references directly in Claude, Cursor, or other MCP-compatible tools.

Official MCP server from Guideline giving AI agents read access to media plan data, campaign details, spend pacing, and performance insights across the Guideline media management platform.

MCP server from Pacvue connecting retail media and commerce advertising data to AI assistants. Enables brands and agencies to pull advertising performance data across Amazon, Walmart, and other retail platforms directly into ChatGPT, Claude, and Copilot.

Agent-native MCP server from SurgeGraph for Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) workflows. Packages research, content generation, publishing, and AI citation monitoring into a single Go binary callable from terminal or AI agent environments. Includes local SQLite mirror for offline queries.

MCP server from Aprimo connecting AI agents to enterprise digital asset management and content operations. Enables approved content to flow into external AI-powered workflows while maintaining governance and brand compliance. Part of Aprimo's Interconnected Content Operations platform linking AI, DAM, work management, and marketing spend.

MCP server from Scenario.com that exposes game art and creative asset generation tools to AI agents and coding assistants. Provides 12 creative tools for generating, editing, and managing game assets including characters, environments, items, and UI elements through MCP-compatible clients. Supports workflow publishing, request tracing, and diagnostics. Enables AI coding agents in Cursor, Claude Code, and similar environments to generate and iterate on visual game assets as part of development workflows without switching to a separate design tool.

Open-source suite of production-validated AI agent skills from Kaltura that enable any AI coding agent to build intelligent rich media applications. Provides MCP-compatible modules for Claude Code and GitHub Copilot to create secure, compliant video and media experiences including avatar-guided portals, interactive video players, live streaming integrations, and content management workflows. Released May 2026 as open source, allowing AI agents to generate full-stack media applications on the Kaltura platform in minutes rather than weeks.

Free open-source MCP server from Kyanite Labs that gives AI agents a real video-editing surface. Provides 85+ tools for programmatic video editing with FFmpeg and new content creation with Hyperframes. Includes guardrails for safe autonomous operation, planning tools for multi-step edits, and support for trimming, concatenating, overlaying, adding audio, applying filters, and generating video from templates. Works with Claude Desktop, Cursor, and other MCP-compatible clients. Python library and CLI also included.

MCP server from Twelve Labs that brings video intelligence capabilities to AI assistants. Enables semantic video search, automatic summarization, Q&A over video content, and embedding generation through a standardized MCP interface. Agents can index videos, find specific scenes using natural language, generate concise summaries, and build RAG workflows that chain search and analysis. Powered by the Marengo embeddings model and Pegasus video language model. Compatible with Claude Desktop, Cursor, Windsurf, Goose, and other MCP clients.

MCP server from Pika that gives AI agents access to multi-model creative content generation capabilities. Connects to Claude Code, Codex, and other MCP-compatible agents to provide tools for text-to-video, image-to-video, video editing, image generation, lip sync, voice cloning, music generation, auto captions, and more. Aggregates models including Seedance, Gemini Image, ChatGPT Images, Pika Video, Kling, MiniMax, Veo 3, Sora, ElevenLabs, and Remotion. Includes ready-to-run creative skills for app launch videos, explainers, UGC ads, podcasts, and brand kits. OAuth authentication via the Pika platform.

Open-source MCP server and Claude Code plugin for ComfyUI that enables AI agents to execute image generation workflows, visualize pipelines, manage models, control VRAM, and explore the ComfyUI node ecosystem. Provides tools for running ComfyUI workflows programmatically, generating images from prompts, managing checkpoint and LoRA models, and monitoring GPU resource usage. Designed for AI-assisted creative workflows where coding agents can orchestrate ComfyUI pipelines directly.

Open-source MCP server that gives AI assistants live, accurate knowledge of Autodesk ShotGrid instances for reliable VFX and animation pipeline code generation. Provides tools to inspect ShotGrid project data, query entity schemas, access shot and asset information, and understand pipeline configurations. Designed for technical directors and pipeline engineers who want AI agents to generate correct ShotGrid integration code based on the actual state of their production database.

Official MiniMax MCP server for text-to-speech, voice cloning, image generation, and video generation. Supports Claude Desktop, Cursor, and other MCP clients with multimodal AI generation capabilities.

Guardrailed video editing MCP server powered by FFmpeg. Supports media analysis, subtitle generation, audio processing, effects, Hyperframes rendering, and repurposing tools. Local, fast, and free.

Suite of three MCP-accessible AI agents from Displayce for digital out-of-home advertising. Transforms DOOH campaign workflows from brief to post-campaign analysis. First agent converts natural language briefs into actionable media plans with programmatically available screen selections. Second agent generates editable sales presentations from audience and screen data. Third agent produces narrative performance reports. Backed by 10+ years of programmatic DOOH data across 2M+ screens in 80 countries. JCDecaux majority-owned.

Open-source MCP server that gives AI agents full control over DaVinci Resolve Studio through the official Scripting API. Provides comprehensive tools for video editing, color grading, media pool organization, render queue management, Fusion compositing, Fairlight audio, and project lifecycle operations. Includes a local browser control panel for inspecting Resolve state and running source-safe media analysis. AI agents can browse projects, swap clips, apply color grades, set up renders, and analyze frames with AI vision. Compatible with Claude, Cursor, and other MCP clients.

MCP server from Bitmovin that enables AI agents to perform automated video playback tests and analyze results in real time through the Stream Lab testing platform. AI agents can create and manage test setups, run tests, and analyze playback on over 30 physical test targets including Samsung, LG, and Vizio TVs, web browsers, and gaming consoles using natural language commands. Built in partnership with Alpic. Launched at CES 2026. Compatible with Claude, Codex, Cursor, Windsurf, GitHub Copilot, and ChatGPT.

MCP server from Bitmovin that provides natural language access to video playback metrics, license usage, and viewer experience insights from the Bitmovin Observability analytics platform. AI agents can explore Quality of Experience data, query performance metrics, and render interactive charts and visualizations directly alongside queries. Pairs with the Stream Lab MCP Server for AI-driven testing guided by real-world QoE data. Compatible with Claude, Codex, Cursor, Windsurf, GitHub Copilot, ChatGPT, and other MCP clients.

MCP server from Flato that connects AI agents to an editable canvas runtime for creating and managing interactive presentations, slide decks, infographics, and visual content. Gives AI full control over text, styles, layout, images, charts, vector elements, animation, and interaction on a persistent canvas that both AI agents and human editors can collaborate on. Includes built-in multi-source image search, vector element generation, and programmable charts. Compatible with Claude Code, Cursor, Codex, and other MCP clients.

MCP server from Pictory for AI-powered video creation workflows. Exposes modular tools including create-storyboard and render-video that handle input validation, sequencing, and API orchestration. Enables AI assistants to generate personalized videos at scale by combining CRM data with video templates, voiceovers, and branding. Fully compliant with the Model Context Protocol for discovery and orchestration through natural language. Works with Claude Desktop, Cursor, and other MCP-compatible automation tools. Plans to expand with transcription, summarization, and PowerPoint input support.

MCP server from Wistia that connects a video library to Claude, Cursor, or ChatGPT for full video management through natural language. Enables creating, searching, editing, publishing, and pulling analytics for an entire video library without opening the Wistia dashboard. Key capabilities include finding content by description, reading captions and transcripts at scale, auditing accessibility gaps, splitting long videos into chapters, curating channels, and performance reporting. Over 300 accounts connected within days of launch generating thousands of queries. Available to account owners and managers. Launched July 6, 2026.

MCP server from Verbatik Technologies that connects text-to-speech voice generation tools directly to AI assistants. Provides access to over 2,700 neural voices in 50-plus languages, instant voice cloning from a 10-second audio sample, and SSML support for fine-grained speech control. Exposes nine tools covering voice library browsing, text-to-speech generation for up to 50,000 characters, cloned-voice generation with seven emotional settings, voice management, account balance checks, and cost estimation. Supports multi-step audio workflows where an assistant can translate a script into multiple languages and generate narration for each version in a single pass. Uses Streamable HTTP transport with OAuth 2.1 and PKCE or per-client API key authentication. Serves over 150,000 users. Free initial setup with complimentary starting credits. Announced July 16, 2026.

MCP server from Revid.ai that gives AI agents direct access to a complete video production pipeline. Exposes tools with stable names including render_video, get_project_status, export_video, clone_voice, schedule_publish, and publish_now. Supports nine production workflows at launch covering script-to-video, prompt-to-video, and audio-to-video from source materials including prompts, scripts, links, and audio files. Produces complete short films with scripts, voices, captions, and correct aspect ratios in a single pass rather than isolated clips. Users receive editable projects with full commercial rights and no watermarks. Supports OAuth 2.1 for hosted agent clients and API-key authentication for local configurations. Based in Paris. Launched July 16, 2026.

MCP server from ngram that enables AI agents to generate professional product demo videos directly from code context. Connects to Claude, Codex, and other MCP-compatible clients to create product videos using specs, release notes, documentation, screenshots, and application state as inputs. Supports the full video production workflow from prompt to rendered output including approval, rendering, progress tracking, review, and QA steps. Designed for developer marketing and product teams who want to automate demo video creation as part of their release process. Updated July 16, 2026.

MCP server platform from Pixazo that provides plug-and-play access to leading AI image and video generation models natively inside any MCP-compatible AI assistant. Offers per-model MCP servers for Nano Banana (Google image generation), Seedance (ByteDance video), Runway (cinematic video), and Flux (Black Forest Labs image generation), plus additional servers for audio, music, 3D, upscaling, and virtual try-on. Developers authenticate once with OAuth through a Pixazo account and use unified credits across all models. Each model runs as a dedicated server for tighter control over outputs and predictable latency. Launched July 16, 2026.

MCP server built into Boris FX Silhouette, the professional VFX and post-production compositing application. Lets AI assistants work directly inside Silhouette projects using the FX Scripting API as first-class MCP tools. Assistants can inspect projects, build node trees, edit shapes and keyframes, and render frames. Supports interactive online mode connected to an active session and offline headless mode for automation, batch processing, and large-scale workflows. A new preferences panel simplifies setup by installing the MCP package, generating client configuration, and testing the connection. Announced at SIGGRAPH July 20, 2026 alongside NVIDIA agentic AI creative ecosystem updates.

MCP-based AI Connector from Affinity by Canva that brings natural-language automation directly into the Affinity design suite (Designer, Photo, Publisher). Lets Claude handle repetitive production tasks such as renaming layers and artboards, resizing and reformatting assets for multiple channels, applying bulk edits, optimizing vector paths and preparing files for delivery. Also enables users to build reusable scripts and custom features tailored to their workflows. Announced at SIGGRAPH July 20, 2026 as part of NVIDIA's agentic AI creative ecosystem updates.

MCP support built into SideFX Houdini 22 through the new APEX Script workflow. AI assistants can access a curated collection of APEX Script syntax, functions, documentation and examples, helping artists generate and refine code for procedural character rigs. Initial implementation focuses on APEX Script and character rigging, while community-developed MCP servers offer broader ways for agents to interact with Houdini. Announced at SIGGRAPH July 20, 2026 as part of NVIDIA's agentic AI creative ecosystem updates.

MCP server from Masset, the digital asset management platform for business content, that exposes 32 tools for AI agents to search, manage and organize content assets. Provides read and write access to DAM operations including asset search, metadata management, content organization and media operations. Designed for marketing teams, content operations and enterprise knowledge management. Enables AI assistants to find and work with brand assets, documents and media files through natural language. Native MCP integration distinguishes Masset from competitors in the DAM space.

MCP server built into the Violet Audio dMix 128 digital mixing console that enables AI-powered mixing and workflow automation. Allows AI assistants to control mixing parameters, manage audio routing and automate repetitive mixing tasks through natural language. Bridges professional audio engineering with AI agent workflows. Part of Violet Audio's push to bring AI directly into professional audio hardware. Announced July 25, 2026.

Open-source guardrailed video editing MCP server from Kyanite Labs that gives AI agents control over video editing operations. Built on FFmpeg with Hyperframes support for frame-accurate editing. Includes repurposing tools for content adaptation, a Python client library and CLI. Runs locally for fast processing. Designed with safety guardrails to prevent destructive operations on source media. Free and open source on GitHub.

MCP server from Getty Images that gives developers and enterprise technology teams a standardized way to integrate licensed creative, editorial and archival visual content into AI-powered workflows and products. AI agents can discover, retrieve and use images and video from Getty Images' collection spanning creative, news, sport, entertainment and archival content. Designed for enterprise customers, media organizations and technology partners building AI solutions for marketing, advertising, news and entertainment. Announced August 12, 2026.

The most capable MCP server for Final Cut Pro with 88 tools covering three layers: FCPXML engine with 49 tools for timeline editing, live FCP control via AppleScript with 20 tools, media analysis via ffprobe with 10 tools, plus parametric puppets with 7 tools and Compressor dispatch with 2 tools. Enables AI assistants to analyze timelines, perform QC validation, edit clips, generate rough cuts, sync beats and import SRT transcripts. Works with Claude, Cursor and other MCP-compatible clients on macOS.

Official local MCP server for the snapWONDERS API that lets an AI assistant work with files on the user's own machine. Supports hiding files inside images, revealing hidden content, forensic analysis, media analysis and file-format conversion by reading local file paths, uploading to snapWONDERS over HTTPS and writing results back to disk. Requires a snapWONDERS API key. Public GitHub repository created August 16, 2026.

MCP server that lets AI assistants control FL Studio through a FastMCP stdio server paired with an FL Studio controller script. Exposes transport, mixer, channel rack, plugins, patterns, step sequencer, piano roll, playlist, project metadata, markers and raw FL scripting API access over a virtual MIDI trigger plus JSON command channel. Public GitHub repository created August 15, 2026.

Small cross-platform MCP server that lets AI assistants make the local computer speak through ElevenLabs text-to-speech, with optional system volume set and restore behavior on macOS, Windows and Linux. It supports stdio for local clients, optional HTTP for remote connector scenarios, and a non-MCP Python library or CLI path for direct gateway integrations. Public GitHub repository created August 15, 2026.

Production-ready unofficial MCP server for the Pexels API, giving assistants tools to search and browse stock photos and videos with compact JSON results and license-aware usage guidance. It is published as an npm package for zero-clone npx use across Claude Desktop, Claude Code, Cursor, VS Code, Windsurf and generic stdio MCP clients. Public GitHub repository created August 11, 2026.

Local MCP server for inspecting, creating and editing Figma Desktop designs through an authenticated localhost plugin bridge instead of a cloud token. It exposes focused tools for screenshots, design specs, JSX rendering, tokens, assets, components, FigJam and Figma Slides, with optional REST support for version history and comments. Public GitHub repository created August 18, 2026.

Local stdio MCP server for Katto that turns long videos into scored, captioned 9:16 clips through the Katto REST API. Tools create clip jobs from YouTube, Twitch, Vimeo, Rumble, Zoom and other video URLs, poll status, list jobs, fetch generated clips, retrieve transcripts and report usage quota. Public GitHub repository created August 20, 2026.

CLI and MCP server for making AI-generated 3D assets web-ready. It analyzes GLB assets against performance budgets, optimizes meshes and textures, builds LOD chains, scaffolds React Three Fiber viewers, generates or downloads Meshy assets, extrudes logos into 3D and exposes these workflows through @glbforge/mcp tools. Public GitHub repository created August 20, 2026.

Hosted HTTP MCP server for paid social advertising automation. Pixero researches a brand site, generates ad creative, structures Meta campaigns and prepares launch/optimization workflows so an agent can make ads rather than only report on existing campaigns. Public GitHub repository created August 14, 2026.

Rust MCP server and local mirror for Figma files that acts as a high-performance superset of the official Figma MCP surface. It syncs Figma data into a local, instantly queryable cache, exposes read/search/dump/image tools, and can proxy the official Figma desktop Dev Mode tools while avoiding repeated API-rate-limit hits. Public GitHub repository created August 17, 2026.

Remote stock-photo search MCP server for AI assistants. It exposes semantic text search, visual search by image URL and similar-photo lookup, renders result grids inline, uses OAuth 2.1 by default with API-key fallback, and requires no local install. Public GitHub repository created August 16, 2026.

MCP server for agentic game-asset generation, taking prompts or reference images through image, 3D model, PBR texture and provenance workflows. It supports new asset generation plus retexturing existing owned meshes, records provider task IDs and SHA-256 hashes, and can normalize meshes with optional Blender support. Public GitHub repository created August 24, 2026.

Official MCP server for UnrealUGC that lets MCP assistants browse creators and models, estimate credit costs and generate AI UGC video ads. It supports local stdio via npm and a hosted streamable HTTP endpoint for clients that can connect remotely. Public GitHub repository created August 18, 2026.

MCP server that engraves GNU LilyPond source into cropped, placeable PDF, EPS, SVG and PNG music assets. It exposes tools for engraving files or inline code and returns LilyPond diagnostics with line numbers so agents can iterate on notation and produce assets for InDesign or other layout software. Public GitHub repository created August 24, 2026.

Built-in MCP server for Doop, an open-source multiplayer design canvas where humans and AI agents design together live. Agents connect over HTTP with OAuth, create and edit frames, stream design work into shared canvases, set status, and inherit the same access boundaries as the human who authorized them. Public GitHub repository created August 22, 2026.

Hosted Streamable HTTP MCP server from HasData for read-only public Instagram data. It lets agents look up public profiles by handle and walk public post feeds as structured JSON without signing in, posting, or acting as an Instagram account. Public GitHub repository created August 21, 2026.

Hosted Streamable HTTP MCP server from HasData for read-only public TikTok data. It exposes tools for public profile lookup, account video feeds, video comments and TikTok keyword search, returning structured JSON without OAuth or a TikTok developer account. Public GitHub repository created August 24, 2026.

MCP bridge that lets AI assistants drive Adobe Photoshop, After Effects, Premiere Pro, Illustrator, Audition and Media Encoder. It advertises more than 120 creative-production tools plus an ffmpeg lane, with per-application panels and live-run documentation. Public GitHub repository created August 26, 2026.

MCP server for IMSLP, the Petrucci Music Library. It provides no-key read access to works, scores and recordings through the public IMSLP API, making classical-music metadata and score references available to assistants. Public GitHub repository created August 27, 2026.

Production-oriented MCP server for the Canva Connect API, packaged for npx and aimed at Gemini CLI, Antigravity, Cursor and Claude-style clients. It lets agents design, manage and export Canva assets through typed tools with OAuth-backed Canva access. Public GitHub repository created August 27, 2026.

Safety-first local MCP integration for VCV Rack 2.6.6 that discovers running Rack instances, inspects patches, explains and validates signal flow, previews and atomically applies structured patch changes, manages checkpoints, and reads Probe telemetry through an authenticated loopback bridge. Public GitHub repository created September 2, 2026.

Small Python stdio MCP bridge between Claude Code and a live PyMOL session over PyMOLs XML-RPC server. It exposes tools to run PyMOL commands, execute Python scripts inside PyMOL, and query XML-RPC methods for structure fetches, selections, representations, coloring, measurements, and atom or chain lookups. Public GitHub repository created September 2, 2026.

Evidence-grounded YouTube transcription MCP server for finding exact moments in long videos. It uses Gemini transcription alongside captions, stores timestamped passages and matching frames locally, and lets agents answer cross-language video questions with citations to the original timeline instead of unsupported summaries. Public GitHub repository created August 30, 2026.

Early-alpha Python MCP server for AI-driven audio editing in Audacity 4. It connects MCP clients to a locally running Audacity 4 companion fork through a TCP JSON-RPC bridge, exposing local audio editing, effects, VST3, cleanup, and project-control commands without cloud processing or audio API keys. Public GitHub repository created September 5, 2026.

Remote Streamable HTTP MCP server from Arcmira for searching the spoken web. It exposes eight read-only tools over indexed YouTube and podcast transcripts, including full transcript retrieval, show-level search, mention lookup, momentum, sponsor, and coverage analysis, with OAuth for hosts that support MCP authorization or bearer/free-trial keys for other clients. Public GitHub repository created September 8, 2026.

Open-source MCP-driven video-generation pipeline for product demo and launch videos. It combines Remotion and Playwright with an npx-installable MCP server so Claude Code, Cursor, ChatGPT, or another MCP client can drive browser capture, keyframing, narration, music-bed, and rendering workflows through sixteen tools. Public GitHub repository created September 2, 2026 and actively updated September 8, 2026.

Early-alpha local Streamable HTTP MCP bridge that gives an AI companion its own Mineflayer-backed player in a Minecraft Java LAN world. It exposes explicit Minecraft action tools through a localhost bearer-token bridge, has been live-tested on Java 1.21.11, and keeps the model subscription outside the server. Public GitHub repository created September 9, 2026.

Streamable HTTP MCP server for Substack designed for Vercel deployment. It exposes twenty-seven tools covering drafts, post body updates, publishing, subscribers, analytics, tags, comments, Notes, and reader feed operations, while keeping Substack publication URL, session cookie, and user ID in server-side environment variables. Public GitHub repository created September 10, 2026.

Python MCP toolkit for ESPN fantasy-football draft and season operations. It includes a manager MCP server, an ESPN MCP companion, and a portfolio MCP for configured teams, with tools for league analysis, lineup planning, browser-assisted drafts, authenticated HTTP season observations, saved policy, transaction leases, and optional controlled roster actions. Public GitHub repository created September 8, 2026.

Windows stdio MCP server for MikuMikuDance authoring and MME configuration. It can inspect MMD windows, models, bones, camera, lights, UI state, timeline frames, expressions, accessories, physics, EMM settings, and AVI export paths, with native editing support limited to verified MMD builds and locally built helper DLLs. Public GitHub repository created September 8, 2026.

Official MCP server for the Paxa Labs API, focused on Thai and English speech AI plus document OCR. It provides tools for text-to-speech playback, queued long-form speech, playback control, audio-file playback, speech-to-file synthesis, Thai translation, PDF and image OCR, audio transcription with subtitles, model listing, voice listing, and account-credit status. Public GitHub repository created September 5, 2026.

Node.js MCP server for Meta Muse image generation. It supports text-to-image generation, editing from reference images, and multi-turn iterative refinement, returning absolute file paths instead of base64 image payloads and reporting estimated call cost while keeping Muse API credentials in the MCP environment. Public GitHub repository created September 12, 2026.

Python stdio MCP server for AI-driven pixel-art and painting toolchains. It exposes gated tool families for Retro Diffusion generation, Aseprite workbench operations, Krita bridge controls, palettes, animation, tilemaps, exports, filters, selections, and per-layer image interchange. Public GitHub repository created September 14, 2026.

Python Streamable HTTP MCP server and REST API that turns curated ComfyUI workflow JSON files into typed MCP tools and REST endpoints. It supports workflow schemas, default and enum parameters, hot reload, image inputs, inline MCP image outputs, job polling, bearer auth, Kubernetes probes, OCI image packaging, and Helm deployment. Public GitHub repository created September 15, 2026.

HTTP MCP server that lets MCP clients request image generation through the Codex CLI and a signed-in ChatGPT account instead of per-image OpenAI API calls. It runs in Docker, queues one Codex-backed image job at a time, serves generated image URLs from the same container, and deliberately avoids forwarding OPENAI_API_KEY into the Codex process. Public GitHub repository created September 14, 2026.

Remote Cloudflare Workers MCP server that wraps an existing ai-images-pilot Worker through a Cloudflare Service Binding. It exposes health, image listing, semantic image search, image processing, and R2 object listing tools backed by Worker AI, D1, R2, and Vectorize resources, with client access through a deployed /mcp endpoint. Public GitHub repository created September 18, 2026.

Lightweight Python bridge that connects AI coding harnesses to live Autodesk Maya sessions through commandPort-based local sessions. It includes a Maya listener, CLI bridge, PySide2 dockable panel, multi-session port discovery, global or per-project skill installation, and guarded userSetup.py patching. Public GitHub repository created September 17, 2026.

Python MCP server for helping AI agents inspect image dimensions and extract sprite-sheet or collage elements with pixel-level bounding boxes. It provides tools for whole-image metadata, sub-image box detection with transparent, white, black, or automatic background handling, grid-box generation, and crop-and-save workflows. Public GitHub repository created September 19, 2026.

Local Node stdio MCP server for reading and editing Compositor .comp design projects through their documented file format. It does not need the Compositor app running to modify a project; inspect changes in the app before trusting generated design edits. Public repository created September 22, 2026.

Local stdio MCP server that validates MapLibre style specifications, renders and compares map styles, and can show interactive maps in MCP Apps-capable clients. Rendering needs Chrome or Chromium; core setup does not require an API key. Public repository created September 24, 2026.

Local Go stdio MCP server that makes pictures and music in headless Mario Paint via a SNES emulator core. Users supply their own compatible emulator core and lawful ROM; the repository builds a separate MCP binary and direct CLI utilities. Public repository created September 23, 2026.

MCP server hosted inside a TouchDesigner component for inspecting and controlling a running project from coding agents over local streamable HTTP. Install the released .tox component in TouchDesigner and enable it; tool calls can mutate the open project. Repository created September 21, 2026.

Local-first Ableton Live 12 bridge with a Remote Script, Node stdio MCP server and CLI for reading tracks, clips and devices and planning guarded changes. Live Set mutations require a single-use confirmation; macOS, Node 22.13+ and a configured Live control surface are required. Repository created September 23, 2026.

Desktop AI video editor with an MCP integration for finding footage, editing clips on a multitrack timeline, captions and export. The server runs only while the Daydream app is open; footage is described by the provider as local-first. Repository created September 24, 2026.

Self-hosted Railway HTTP MCP server that inspects a screen recording and uses FFmpeg plus ElevenLabs speech to create a narrated tutorial, MP4, script and thumbnail. Each operator must deploy their own service and key; its MCP endpoint has no user authentication, so restrict access before exposing private recordings or paid voice generation. Repository created September 24, 2026.

Local Python MCP server that samples video frames with timecodes and reports container metadata and optional scene, silence and loudness analysis using FFmpeg. The client model must support image input to view returned frames. Repository created September 24, 2026.

Claude Desktop MCP extension and Lua bridge for DaVinci Resolve 21.1 free edition, exposing project, media pool, timeline, marker and render tools plus arbitrary Lua execution. The bridge runs only after a user launches it inside Resolve; Windows support is experimental. Repository created September 21, 2026.

Community stdio MCP server for PreSonus Studio One that reads saved .song archives and controls a running session through an installed control-surface bridge. Offers track, clip, tempo, marker, mixer and recording tools, including destructive edits; the author calls it early and tested primarily on Studio One 5.5.2 on macOS. Repository created September 28, 2026.

Unofficial local MCP and CLI for controlling Google Flow via a logged-in Chrome profile: inspect projects and credits, generate Veo video or images, and assemble scenes. The project documents a spend ledger and double-charge guard, but paid actions still require explicit review; tested on macOS with a migrated Flow account. Repository created September 27, 2026.

Windows-local MCP bridge between an agent and a running WARUDO scene, with tools for characters, poses, expressions, cameras, assets, blueprints and captures. The bridge communicates locally through files and is not affiliated with WARUDO or HakuyaLabs; review scene mutations before execution. Repository created September 28, 2026.

Local agent-oriented digital audio workstation with an MCP interface for composing music as notes and code, rendering tracks, and reading back audio analyses as text. It also exposes a CLI and optional live playback; it is not a prompt-to-song model. Repository created September 26, 2026.

Unofficial macOS MCP integration for recording and editing Screen Studio projects, including cuts, zooms, masks, narration and export. Editing compatibility is tested against Screen Studio 4.0.1; other builds require explicit opt-in. Screen control and export are side effects. Repository created October 2, 2026.