claudeers.
// Uncategorized / Others

dramaclaw

A general-purpose AIGC video engine: script to finished film in one pipeline — dramas, ads, product videos, otome games, and more. | 通用 AIGC 视频引擎 —— 从剧本到成片一条…

Install with your AI

Paste into Claude Code, Cursor, or any agent — it reads the repo and wires the tool into your project.

Install and set up dramaclaw (git-clone project) into my current project.
Found on https://claudeers.com/dramaclaw
Repo: https://github.com/dramaclaw/dramaclaw
Homepage/docs: https://dramaclaw.ai
Detected install method: git-clone → git clone https://github.com/dramaclaw/dramaclaw
Category: uncategorized. Platforms: api, desktop, web.
Read the repo's README for exact setup and env vars, then install it and wire it into my project.

Claudeers Health Verdict:
unknown; community-verified: false. Confirm the source before running anything.
// or clone
git clone https://github.com/dramaclaw/dramaclaw

// compatibility

Platformsapi, desktop, web
Operating systems—
AI compatibilityclaude
LicenseNOASSERTION
Pricingopen-source
LanguageTypeScript

Get your FREE $2.50 API credits to access TickAtlas financial data ↗

DramaClaw

Make Your Own DC.

They say you're obsolete.
Maybe it's the whole idea of "working for someone else" that's obsolete.

In the age of AI, the real question isn't whether machines replace people.
The real question is:
Who owns the machines?
Who owns the pipeline?
Who owns industrialized productivity?

If the answer is always Big Tech,
then AI isn't empowerment.
It's just a new wall.

I'm Eric.

This isn't a demo.
Not a toy.
Not a crippled edition.

This is the industrialized drama-production line our own team runs every day.
From script to storyboard, from assets to finished film — the whole chain.

Because people aren't beasts of burden.
Because creativity is humanity's last line of defense.

What DramaClaw sets out to do is simple:

Tear down the wall.

Put the industrialized drama-production power that only Big Tech had
into the hands of ordinary creators.

The code is here.
If this resonates, leave a ⭐.
We'll keep tearing down walls.

And yes, you can make money with it.
DramaClaw is Elastic License 2.0. Run it, modify it, sell what you build with it —
the standalone version is free for commercial use, no permission needed.
We ask for one thing: a small "Powered by DramaClaw" in the corner of your UI.
The only door still closed is wrapping it up as a hosted SaaS for other people; that license isn't open yet.
Why, and the FAQ: #475.


English  |  简体中文  |  Tiếng Việt  |  ไทย  |  Website  |  Docs  |  Quick Start


DramaClaw — storytellers, back in front of the camera

🎬 Trailer: Bilibili  ·  YouTube


What is DramaClaw?

DramaClaw is a self-hostable production line for AI drama: from manuscript to finished episodes, and from a one-line brief to a short film. An agent and an infinite canvas share one asset library:

  • Xia Director — the built-in agent. Describe what you want; it plans the workflow, lays out the nodes and runs them with your confirmation.
  • XiaHua — the infinite canvas. Generate, edit and connect images, video, audio, 3D sets, interactive stories and scripts as nodes.
  • Series (XiaJi) — the pipeline. Drop in a novel or screenplay; DramaClaw structures the story, writes scripts, draws storyboards, voices the lines and cuts the film.

Bring any model through an OpenAI-compatible gateway, run it on your own machine, and keep your material.

🚀 Quick Start  ·  🆕 What's New  ·  🎬 Showcase  ·  ✨ Features  ·  🔌 Models  ·  📖 Docs


🆕 What's New

Three things we have been building with our own production team, now in main.

🤖 Xia Director — from a brief to a running workflow

Xia Director turns a natural-language brief into a canvas workflow

One brief → planned workflow → nodes on the canvas → generated assets → composed video.

  • Plans and runs canvas workflows — turns a creative brief into a node graph, lays it out, and executes it after you confirm, with progress tracking and failure recovery.
  • Workflow catalog, Skills and Recipes — built-in workflows for short drama, e-commerce and IP character ads, social content, Japanese anime, LEGO-style animation and outdoor stage duels. Install, export and share Skill bundles, or import an external SKILL.md / ZIP and convert it into a native Skill.
  • Builds web pages too — ask for a landing page or story microsite; it becomes an editable HTML node with revision history and a self-contained ZIP export.
  • Knows your project — checks progress, advances script and shot tasks, finds missing deliverables and suggests what to do next.
  • Your choice of brain — ships with a bundled Codex runtime by default; point the Agent at any third-party model that supports the OpenAI Responses API and tool calls, independent of your media channels.
  • Works from other agents — Agent Kit exposes the same workflows over MCP to Codex, Claude Code, OpenClaw, WorkBuddy and any stdio MCP client.

🎥 3D Previz Stage — block the scene before you generate it

3D Previz stage: blocking, camera and timeline, recorded back into the canvas

Reference image → 3D blockout → characters, cameras and timeline → recorded preview back on the canvas.

An in-browser previs stage that lives on the canvas as a node. Lock down space, blocking and framing first, then feed the result to image and video generation.

  • Scene blockout — generate a 3D scene from a reference image; drop props and they land on whatever surface is underneath.
  • Characters and motion — mannequins in five builds, placed from a top-down map; walk cycles that follow movement paths on a multi-track timeline.
  • Cameras and cutting — focal length, sensor size and placement per camera; a program track to cut between cameras (digit keys cut live) with the monitor following the edit.
  • Audio track — import audio from a local file or an upstream node, see waveforms, and mix it into recordings.
  • Back to the canvas — capture frames or record the program cut into connected nodes as references for the next generation.

🔀 Interactive Stories — audience choices shape the story

Interactive stories: branching authoring and player

Branching story generation, canvas authoring, interactive playback and multiple endings.

  • AI story planning — a brief becomes an outline and an editable branching narrative with choice points, endings and estimated runtime per route.
  • Visual branching on the canvas — arrange segments, videos and choices; edit dialogue, choice order and branch connections.
  • Variables and conditions — affinity, scores and other variables gate choices and decide endings; timed choices and default routes included.
  • Produce media in place — character, scene and storyboard references and per-segment videos, reusing every canvas asset.
  • Playtest before rendering — run the story with placeholder media, validate branch connectivity and conditional paths, inspect ending reach rates and path coverage.
  • Ship a player — export a standalone HTML player or an offline package with video, branching playback, progress and replay.

🎬 Made with DramaClaw

Real short dramas our team produced on this very pipeline — click a link to play.

归灵司
归灵司
▶ 第 1 集
鲁班
鲁班
▶ 第 1 集
师兄你怎么不舔了
师兄你怎么不舔了
▶ 第 1 集
天命不可欺
天命不可欺
▶ 第 02 集  ·  第 58 集
乌龙仙途
乌龙仙途
▶ 第 1 集
非遗㑇舞
非遗㑇舞
▶ 播放

More clips: 3D 动漫混剪 demo  ·  动态打斗


🚀 Quick Start

You need Docker with docker compose. No GPU, database or queue is required — models run behind the gateway, state lives on local disk.

git clone https://github.com/dramaclaw/dramaclaw.git
cd dramaclaw
cp .env.example .env          # set PROMPT_EXPORT_PASSWORD to a non-default value

docker compose -f docker-compose.release.yml pull
docker compose -f docker-compose.release.yml up -d

Open http://localhost:8080, go to Settings → Models & Channels, and paste your DC key (get one at https://relayclaw.cdnfg.com) — or switch to Custom to use your own providers. That's it.

The release Compose file runs the published v2.1.1 multi-arch images (DRAMACLAW_VERSION=2.1.1, DRAMACLAW_GATEWAY_VERSION=v1.0.0-rc.24-dramaclaw.3). Mainland China users can switch to the ACR mirror with DRAMACLAW_IMAGE_PREFIX.

Other ways to run it
MethodWhen to useCommandGuide
Published imagesTry or self-host a releasedocker compose -f docker-compose.release.yml up -dQuick Start
Source DockerRun your own modified codeclone dramaclaw-gateway next to this repo, then docker compose up -d --buildSelf-hosting
Local devHack on DramaClaw with hot reloadscripts/start-ce.sh (API :8780 + UI :5173; needs Python 3.11–3.12, uv, Node/pnpm, ffmpeg)Installation

scripts/start-ce.sh downloads the patched Codex 0.164.0 Agent runtime for macOS or Linux, verifies its pinned SHA256, and caches it under runtime/codex/bin/. Set CODEX_BIN in .env to use your own patched build instead; Windows users should use WSL2 or Docker.

Upgrading from an older checkout, pinning versions, exposing the bundled gateway, or using your own Codex binary: see the Self-hosting guide and Troubleshooting.

System requirements
ItemRequirement
CPU / RAM≥ 2 vCPU / 4 GB (model inference runs on the gateway, not locally)
GPUNot required. Only the optional world extra (voxel / panorama-to-3D) needs CUDA
OSmacOS (Apple Silicon / Intel), Windows (Docker Desktop + WSL2), Linux
Ports8080 web UI · 8780 REST API · 3000 bundled gateway admin (localhost only by default)
DatastoresNone — no Postgres, Redis, Celery or Ray. SQLite + files on the ce-data volume
NetworkOutbound access to relayclaw.cdnfg.com and/or the providers you configure

✨ Features

XiaHua — the infinite canvas

XiaHua canvas: a 17-node short-film workflow that Xia Director laid out and ran from one prompt

One prompt to Xia Director (right) became this 17-node, 32-edge workflow (left): brief → story direction → character and set references → five first frames → five shot videos → voice-over and score → final cut.

  • One canvas, every medium — image, video, audio, script, beat context, storyboard, style, skill, 360° panorama, 3D world, Previz, interactive story, web page and export nodes. Connect them freely; every node keeps its own generation history.
  • Image tools — generate, edit, redraw, outpaint, relight, upscale, multi-view, template edit, reverse-prompt, mark detection, and a style wall of 45 short-drama looks.
  • Video tools — text / image / keyframe to video, omni-reference generation with file, web-link and on-canvas references, video edit, erase, upscale, audio separation, camera templates and shot analysis; Seedance 2.5 in the recommended catalog.
  • Audio and voice design — speech synthesis, voice library, reference audio and music. Describe a voice in words, preview it, and save it to a character slot or My voices (needs a compatible audio model and VOICE_DESIGN_ENABLED=1 with daily limits in .env).
  • Director World — turn an image into a 3D Gaussian Splat set or a scene-360 panorama, frame the shot in the viewer, and reuse the same location consistently across shots.
  • Built for big canvases — tabs, outline, minimap, viewport bookmarks, snap-align, level-of-detail rendering, group nodes, shortcuts, revision history with restore and per-canvas locking.
  • Commit to the series — preview the impact, then promote a node or a batch into the asset library or an episode; project the series back onto a canvas.
Director World

Director World

Previz node connected to its reference image

Previz node on the canvas

Series (XiaJi) — the pipeline

DramaClaw pipeline — Ingest, Plan, Produce, Deliver

Every stage is an independent task with its own interface — run them in order, skip steps, resume from any checkpoint, or drive them from your own orchestrator.

  • Structured ingest — episodes, characters and scenes straight from a manuscript or screenplay (Fountain supported); no knowledge graph or embeddings required.
  • Asset library (XiaTang) — characters, scenes, props and voices organized by purpose; stable identities across episodes, portraits and per-episode variants.
  • Episode planning and scripts — chapter segmentation, beat planning, multi-episode arcs; adaptive / literal / staged script modes with review-and-repair loops.
  • Storyboards and first frames — beat-driven stylized generation, grid splitting, image-pool selection.
  • Voice-over, composition and export — emotion-aware speech, episode assembly, video + subtitle export and the full asset pack.
  • Visual style — extract style parameters from a reference image and apply them project-wide.
  • Task Center — status, progress, logs, cancel / retry and resume-from-checkpoint for long runs.
XiaTang asset library

XiaTang — asset library

Storyboards and first frames

Storyboards and first frames


🤔 Why DramaClaw?

  • A canvas that understands drama, and an agent that works on it. Generic node canvases wire anything together but know nothing about episodes, characters or shots. DramaClaw's nodes and Skills do, and Xia Director plans and runs them for you.
  • Canvas and pipeline. Explore freely on the canvas, commit what works to the series, project the series back onto a canvas — one asset library, one agent.
  • No black box. Every stage is a decomposable task with its own API. Skip, resume, replace or script any step.
  • Self-hostable and model-neutral. Your manuscript, your characters, your models, your servers. Use frontier closed models or open-weight ones; switch any time.

How DramaClaw compares

Legend: ✅ Full · ◐ Partial · ○ Planned · ❌ None · — Not verified. Competitor names partially masked; based on publicly available product docs and positioning.

CapabilityL*TVR*HubT*NowS*koU*dreamO*IIJ*/K*DramaClaw
Built-in agent (workflow planning, canvas orchestration, execution)✅✅✅○✅○✅✅
Workflow Skills and Recipes (reuse, import, export, share)———————✅
3D previz stage (blocking, motion, cameras, timeline, recording)———————✅
Interactive series (branching, variables, multiple endings)◐◐◐✅◐❌❌✅
Infinite canvas (node-based, free exploration)✅✅✅❌✅✅✅✅
Dual-track (main pipeline + canvas exploration)❌❌❌❌❌❌❌✅
Storyboard preview (script → shots, boards)◐✅◐✅◐❌❌✅
Asset library (characters / scenes / props / voices)◐❌◐✅◐○❌✅
Scene consistency (variants, 360°, multi-state)✅◐❌❌○❌❌✅
Director's world (360° / 3D set, camera, framing)✅◐❌❌◐❌❌✅
Final delivery (multi-shot, subtitles / SRT, pack)✅○○✅✅❌❌✅
Team production (sharing, roles, tasks, cost)✅✅○✅○○○✅
Custom style (templates, prompts, negatives)✅◐○✅◐○◐✅
Creative companion (persona, nudges, feedback)❌❌❌❌❌❌❌✅
Source-available (self-host, fork, customize)❌❌❌❌❌❌❌✅

🔌 Models & Providers

All text, image, video and audio models connect through an OpenAI-compatible gateway. Pick a mode in Settings → Models & Channels:

ModeWhat you doBest for
Official (recommended)Paste your DC key from RelayClaw. Every model mapping is preconfigured.Getting started fast
CustomOne click initializes the bundled dramaclaw-gateway; add your own provider channels.Full control over providers and cost
Local + Official HybridOfficial models for the pipeline, plus your own channels (e.g. a local ComfyUI workflow).Mixing in local or niche models

The bundled gateway is DramaClaw's fork of New API. It speaks the DC-Media contract and ships adapters for ComfyUI · MiniMax / Hailuo · VolcEngine Doubao / Seedance · fal.ai · Alibaba · Kling · Jimeng · Vertex AI · Gemini · OpenAI / Sora · Suno (support matrix, add a provider).

The Agent model is configured separately under Models & Channels → Agent model and can follow the active channel or use any endpoint that supports the OpenAI Responses API with tool calls. Details: Configuring models.


📖 Documentation

StartQuick Start · Installation · Configuring models
LearnFeature overview · Architecture · Product manual (Feishu)
OperateSelf-hosting · Troubleshooting · Use from Claude Code (MCP)
MoreAll docs →

🤝 Community & Contributing


Produced By

Neo Flying AI Laboratory

Logo is a trademark of its respective owner, shown with permission.


License

Elastic License 2.0. Free to use, modify, redistribute and sell what you build with it; keep a small "Powered by DramaClaw" in your UI. The only restriction is that you may not offer the software itself as a hosted service to others. See the license explainer, the licensing statement, and NOTICE for required branding and third-party attributions.


Star History

Star History Chart



Built for storytellers. Source, open to all.

// faq

What is dramaclaw?

A general-purpose AIGC video engine: script to finished film in one pipeline — dramas, ads, product videos, otome games, and more. | 通用 AIGC 视频引擎 —— 从剧本到成片一条流水线,漫剧、广告、电商、乙游皆可. It is open-source on GitHub.

Is dramaclaw free to use?

dramaclaw is open-source under the NOASSERTION license, so it is free to use.

What category does dramaclaw belong to?

dramaclaw is listed under uncategorized in the Claudeers registry of Claude-compatible tools.

7 views
★ 7,709 stars
unclaimed
updated 4 days ago

// embed badge

dramaclaw on Claudeers
[![Claudeers](https://claudeers.com/api/badge/dramaclaw.svg)](https://claudeers.com/dramaclaw)

// retro hit counter

dramaclaw hit counter
[![Hits](https://claudeers.com/api/counter/dramaclaw.svg)](https://claudeers.com/dramaclaw)

// reviews

// guestbook

0/500

// related in Uncategorized / Others

🔓

Reverse engineer anything with agents, from app behavior down to native binaries.

// uncategorizedmorluto/⟨TypeScript⟩★ 423,040◷ MIT[ claude ]
🔓

Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.

// uncategorizedn8n-io/⟨TypeScript⟩★ 209,482◷ NOASSERTION[ claude ]
🔓

The agent engineering platform.

// uncategorizedlangchain-ai/⟨Python⟩★ 148,973◷ MIT[ claude ]
🔓

100+ AI Agent & RAG apps you can actually run — clone, customize, ship.

// uncategorizedShubhamsaboo/⟨Python⟩★ 144,747◷ Apache-2.0[ claude ]

// built by

→ see how dramaclaw connects across the ecosystem