3dBlock Studio
Generates narration with word-level timing, sound effects and an original score through the ElevenLabs API (voice, SFX and music), caching each line by hash so unchanged lines cost nothing; configured with an ELEVENLABS_API_KEY (or a key file) and an output format setting, and without a key renders fall back to silent drafts. Includes a list_voices tool to browse the user's ElevenLabs voices and a voice_lines tool to voice every line of an episode's spec.
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., "@3dBlock Studio Make a PS1 horror short about a cursed videotape and show me the script first."
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.
3dBlock Studio
An MCP server that lets any AI make PS1-style low-poly 3D short videos. Give your AI a story; it writes the script, voices it with ElevenLabs, builds the scene in three.js, and renders a vertical 1080x1920 video that loops perfectly. It ships with the engines, a prop library, an 11-role AI film crew and the render pipeline behind the Guess The Horror and Strange History channels.
What it makes
PS1 low-poly 3D: box people, tiny painted textures, vertex wobble, fog, per-vertex lighting, then real lenses (mm focal lengths, depth of field), bloom, flares, letterbox and colour grades on top.
Living woodcut broadsheets for history stories.
Narration with word-level timing, sound effects and an original score from ElevenLabs, mixed with the voice always on top and normalized to -14 LUFS.
A seamless loop: the last frame is the first frame, and the audio tail crossfades onto the start.
Related MCP server: ClipTalk MCP Server
Install
You need Node 18+ and Google Chrome (or Chromium/Edge). ffmpeg comes bundled. The renders run on your computer.
Claude Desktop / Claude Code / Cursor / any MCP client
{
"mcpServers": {
"3dblock-studio": {
"command": "npx",
"args": ["-y", "3dblock-studio-mcp"],
"env": { "ELEVENLABS_API_KEY": "your key (optional: without it, renders are silent drafts)" }
}
}
}Claude Code in one line: claude mcp add 3dblock-studio -- npx -y 3dblock-studio-mcp
From a clone: npm install, then point the client at node /path/to/3dblock-studio-mcp/bin/3dblock-studio-mcp.mjs.
Setting | Default | What it does |
|
| Episodes, takes, render jobs and the audio cache |
| none | Voice, sound effects, music. Or put the key in |
| auto-detected | Any Chrome, Chromium or Edge |
| bundled | Use your own ffmpeg |
|
|
|
Repository layout (one repo, one engine)
Path | What it is |
| The MCP server ( |
| The AI film crew briefs, episode templates and worked examples the server hands to your AI. |
| The 3D studio (working name): the single engine, the Doll v2 character system and its interactive editor, camera/lights/post, the render pipeline and tools. |
| The engine itself. The MCP server serves this folder to its pages as |
| The shared timeline library and renderer ( |
| Parked sources of the Conspiracy Theories series, kept separate from the studio. |
There is no second copy of the engine: the MCP server and the studio both use Xanadu/studio/lib.
Doll editor: from Xanadu/, run node tools/serve.mjs and open http://localhost:8765/studio/lib/doll/viewer.html. Build and pose a character: body, hair, face, clothes, armor, animations and emotions.
Studio working folders (Episodes/, Uploads/, Support/, Temp/, sound libraries) are git-ignored and live on the author's T7 drive.
Your first episode in five prompts
"Check the studio setup." (
studio_info)"Make a Guess The Horror short about a cursed videotape. Show me the script first." (
make_episodeprompt:new_episode, writer brief,write_file spec.js)"Looks good, voice it and build the scene." (
voice_lines,read_episode the-cursed-tape-ps1,get_engine_guide ps1,write_file scene.js)"Show me stills." (
render_stills,job_statusreturns the contact sheet; the AI fixes what it sees)"Render a take and open it." (
render_take,job_status,open_take, thenpick_takeandpost_kitwhen you like it)
Tools
Group | Tool | What it does |
Plan |
| Studio folder, Chrome, ffmpeg, whether a key is set |
Plan |
| Series formats: engine, target length, rules |
Plan |
| The look, copyright rules, format playbook |
Plan |
| One crew role's brief (director, writer, casting director, production designer, cinematographer, gaffer, script supervisor, editor, sound engineer, colorist, producer) |
Plan |
| How the ps1, woodcut and timeline engines work, with their API |
Write |
| Episode folder from a template that renders as-is |
Write |
| Your episodes and the built-in examples |
Write |
| Save |
Library |
| People, furniture, places and vehicles builders with their options |
Library |
| Render any prop from 4 angles next to a reference person |
Sound |
| Your ElevenLabs voices |
Sound |
| Voice every line (cached by hash: unchanged lines are free), returns lengths |
Shoot |
| 30-frame contact sheet for a quick look |
Shoot |
| The next numbered take: frames, mix, thumbnail, contact sheet, QC |
Post |
| Render progress; finished jobs return their images |
Post |
| Length, loudness, loop seam score, copyright checklist |
Post |
| Mark the FINAL; play a take on screen |
Deliver |
| Caption, up to 5 hashtags, song note, posting time |
Prompts: one per crew role (/director, /colorist, ...) plus make_episode. Resources: the house rules and the example episodes.
The prop library
How it works
The AI only sends tool calls and code. On your machine the server voices the lines (ElevenLabs, cached), serves the scene to headless Chrome, steps time frame by frame (every scene is a pure function of t), pipes the frames to ffmpeg, mixes the audio with ducking and a loop crossfade, and writes the take next to its thumbnail and contact sheet.
episodes/<slug>/spec.js the script, voice, music, sound cues (word-timed)
episodes/<slug>/scene.js the three.js scene on the PS1 or woodcut engine
output/<series>/<NN - Title>/takes/<Title> - Take NN.mp4House rules (built in)
Retellings only, with original art, music and voices: no real posters, stills, clips, logos or theme music. Guessing formats never name the answer. No gore. Made-up brands only. Every take is numbered; only you pick the FINAL.
License
MIT. ElevenLabs usage is billed to your own account.
Available Tools
23 toolsget_crew_briefBRead-only
Get one film-crew role's brief so you can act as that role. Roles: casting-director, cinematographer, colorist, director, editor, gaffer, producer, production-designer, script-supervisor, sound-engineer, writer.
Inputs: role (required).
get_engine_guideBRead-only
How an engine works and its API: ps1 (low-poly 3D), woodcut (living broadsheet), timeline (shared helpers: at/st/en word timing, easing).
Inputs: engine (required).
get_house_rulesARead-only
The house rules every crew role follows: the look, copyright, format, working rules. Read before writing anything.
Inputs: no named parameters.
job_statusARead-only
Progress of a render job. A finished stills job returns its contact sheet; a finished take returns its file, thumbnail and QC.
Inputs: job_id (required).
list_episodesBRead-only
Your episodes, plus the built-in examples you can read.
Inputs: no named parameters.
list_formatsBRead-only
The series formats this studio makes (engine, target length, rules).
Inputs: no named parameters.
list_jobsBRead-only
All render jobs since the server started.
Inputs: no named parameters.
list_propsARead-only
Browse the low-poly prop library (people, furniture, places, vehicles). Without module: the modules and their builders. With module: its full usage guide.
Inputs: module.
list_takesCRead-only
The numbered takes of an episode.
Inputs: slug (required).
list_voicesARead-only
The ElevenLabs voices on your account (needs a key).
Inputs: no named parameters.
new_episodeC
Create an episode folder from a series template (a working placeholder scene that renders as-is).
Inputs: slug (required), title (required), year, engine, +1 more.
open_takeB
Open a take in the computer's video player so the user can watch it.
Inputs: slug (required), take (required).
pick_takeA
Mark a take as the FINAL (only when the user chooses it). Copies it to " - FINAL.mp4" and refreshes the thumbnail.
Inputs: slug (required), take (required).
post_kitC
Write the post kit next to the takes: caption, up to 5 hashtags, song note, posting time. The answer of a guessing format is kept out of the caption.
Inputs: slug (required), song, caption, hashtags.
preview_propBRead-only
Render one prop builder from 4 angles next to a 1.75 m reference person, and return the image.
Inputs: module (required), builder (required), look, height, +2 more.
qc_takeARead-only
Check a take: length against the series target, loudness (-14 LUFS), the loop seam (first vs last frame), and the copyright checklist.
Inputs: slug (required), take (required).
read_episodeARead-only
Read every file of an episode or example (spec.js, scene.js, notes). Use the examples as reference builds.
Inputs: slug (required).
read_fileCRead-only
Read one file of an episode.
Inputs: file (required), slug (required).
render_stillsA
Start a contact-sheet render (30 evenly spaced frames, or chosen times) for a quick look before a full take. Poll job_status; the finished job returns the image.
Inputs: slug (required), draft, times.
render_takeA
Start a full render of the next numbered take: frames, voice, sound effects, music, loudness-normalized mix, thumbnail, contact sheet and QC. Takes minutes; poll job_status.
Inputs: slug (required), draft.
studio_infoARead-only
Check the setup: studio folder, Chrome, ffmpeg, whether an ElevenLabs key is configured, engines and series formats.
Inputs: no named parameters.
voice_linesA
Voice every line of an episode's script with ElevenLabs (cached by hash: unchanged lines cost nothing) and generate its sound effects. Returns each line's length and the episode's approximate length.
Inputs: slug (required).
write_fileA
Write spec.js (script, voice, music, sound cues), scene.js (the three.js scene) or a crew note (.md/.json). JS is syntax-checked and the spec validated before saving; the old version is kept in .history.
Inputs: file (required), slug (required), content (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 23 tools
Most tools have clearly distinct purposes: setup (studio_info, list_formats, get_house_rules), reading (read_episode vs read_file, clearly scoped), rendering (render_stills vs render_take, distinguished by stills/contact sheet vs full take), and delivery (pick_take, open_take, post_kit). A few pairs could trip an agent: list_takes vs list_jobs (takes vs render jobs), and voice_lines bundles both voice and sound-effect generation. The descriptions help resolve these, so boundaries are mostly clear.
Nearly all tools follow a verb_noun or noun_info pattern: list_formats, get_crew_brief, render_take, qc_take, pick_take, open_take. Minor deviations: 'new_episode' instead of create_episode, and 'job_status' as a noun_noun form breaking the verb pattern. All snake_case is consistent, so it reads predictably.
23 tools is at the upper edge but justified by a genuinely multi-stage pipeline: setup, rules/briefs, episode authoring, props, voicing, rendering, QC, and posting. No tool looks redundant enough to remove, though it is slightly heavy.
The full lifecycle is covered: create/read/write episodes, inspect props and voices, voice lines, render stills and takes, poll jobs, QC, pick final, open, and post kit. Minor gaps: no delete/cleanup for episodes or takes, and no explicit way to update episode metadata beyond rewriting spec/scene. These are workaroundable.
Maintenance
Related MCP Connectors
AI editor to build, animate & export layered short-form video projects via one tool catalog.
Plan, render, fix and publish multi-scene AI videos: scripts, storyboards, cast, narration, YouTube.
AI video editing + publishing: turn clips into vertical shorts, post to TikTok/Instagram/YouTube.
On-brand creative studio for AI agents: images, video, audio, and 3D.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceEnables AI assistants to create finished short-form videos with script, voiceover, scenes, captions, and music, delivering an MP4 ready for TikTok, Instagram Reels, and YouTube Shorts.115 npmMIT
- AlicenseNot gradedqualityCmaintenanceEnables AI assistants to generate, edit and render finished, ready-to-post videos from a plain-language prompt, including script, visuals, voiceover, captions and music for TikTok, Reels and YouTube. Also supports remaking reference videos, importing your own media, and reviewing or adjusting existing videos before rendering.MIT
- AlicenseBqualityBmaintenanceEnables AI agents to script, narrate, source assets, and edit a validated NLE-style timeline with narration-anchored timing, then preview, check and render a finished vertical film with layout, loudness and delivery rules enforced. Every edit is atomic and undoable, so agents can iterate on real editor state instead of a fixed template.19PolyForm Noncommercial 1.0.0
- AlicenseAqualityBmaintenanceEnables any MCP-capable AI assistant to turn an idea into a faceless short video by drafting scripts, generating voiceovers, captions, and scenes, pricing and rendering videos, managing series, and publishing to connected social accounts.5435 npmMIT