AutoDesign README

The project's README from github.com/Yaxin9Luo/AutoDesign. As one markdown file: README.md.

English · 简体中文 · 한국어

News

Date Update
2026-08-19 Poster canvas controls: prompt-first templates, aspect ratios, and exact pixel sizes, plus academic presets in the Web UI
2026-08-19 Agent Skills v0.2.0: Poster Agent-first PDF ingestion and white primary canvases across all four Skills
2026-08-18 Poster Skill Agent-first v2: direct PDF curation, revision-bound attempts, and read-only DOM QA
2026-08-17 Standalone Agent Skills for Poster, PPT, Webpage, and Video are now installable
2026-08-15 Added official DeepSeek Harness support for coding agents
2026-08-14 Initial public release

AutoDesign improves the harness around a fixed model and ships editable posters, slides, webpages, and videos

AutoDesign: Meta-Harness Optimization for Long-Horizon Agentic Design

Learn a reusable DesignHarness around fixed models, then turn one paper into an editable poster, slides, webpage, and narrated + captioned video.

Paper · arXiv:2608.13560 ↗   ·   Dataset · PosterBench ↗   ·   Dataset · PosterBench-mini ↗

✦ Explore the AutoDesign story ↗   ·   Open the Demo Page ↗

Demos · Agent Skills · Quickstart · Paper Suite · Methodology · PosterBench · Human Evaluation · Outputs · Citation

AutoDesign for AutoDesign · One Paper → Four Artifacts

These are real outputs, not mockups. AutoDesign turned its own paper into the paper's Figure 2 poster, a 24-slide formal academic talk, a complete editorial research webpage, and a six-minute 1080p conference video.

AutoDesign for AutoDesign poster from Figure 2 of the paper
Poster · AutoDesign
Figure 2 of the paper: an information-dense, editable academic poster made by AutoDesign for itself.
Open full poster PDF ↗ 		Selected slides from the 24-slide formal AutoDesign academic conference talk
Slides · AutoDesign
A complete 24-slide formal academic conference talk made by AutoDesign for itself.
Open full slide deck PDF ↗
Editorial AutoDesign research webpage generated from the AutoDesign paper
Webpage · AutoDesign
An editorial research experience that turns the paper's method, evidence, results, and limitations into an interactive story.
Download full landing page ↗ 		Contact sheet from the six-minute AutoDesign conference video
Video · AutoDesign
A six-minute 1080p conference video introducing Meta-Harness Optimization, DesignHarness, and PosterBench.
Watch MP4 ↗

Agent Skills · Use AutoDesign without the server

Install the standalone Poster, PPT, Webpage, or Video Skill directly in Codex, Claude Code, or DeepSeek Harness. Each Skill carries its own lightweight local harness and keeps editable artifacts, evidence, attempts, and review state in your chosen output directory—no AutoDesign application server required.

Poster is the first Skill to receive the new Agent-first v2 workflow: the host agent can inspect the paper PDF, request deterministic crops, review the source catalog, and repair the correct workflow stage while a read-only browser audit checks both screen and print output. See what changed →

Open the Agent Skills install guide →

Start locally

One-command local launch

Prerequisites: Node.js 22+ and ffmpeg/ffprobe.

curl -fsSL https://designanything.ai/install.sh | bash
autodesign start

The launcher installs under ~/.local/share/autodesign, keeps state under ~/.autodesign, serves the bundled web/dist, and opens a browser. Run autodesign doctor to check the installed runtime. Existing ~/.designanything state is migrated with a compatibility symlink. If the hosted endpoint is unavailable, use the source setup below.

Build from source

Requirements: Python 3.10+, uv, Node.js 22+, npm, and ffmpeg/ffprobe for Video.

1. Install

uv sync
uv run python scripts/install_playwright_browsers.py
cd runtime/video && npm ci --omit=dev
cd ../../web && npm install

Configure provider keys in .env or enter them in the Web UI Settings drawer. Do not replace an existing .env during an update.

2. Launch the workbench

Start the backend:

uv run uvicorn scripts.web_server:app --reload --port 8000

In another terminal, start the frontend:

cd web
npm run dev

Open localhost:5173. Backend health is available at /api/health.

Upload one PDF and choose Paper All-in-One to launch the poster, slides, webpage, and narrated-video tracks together.

3. Generate a paper poster

uv --cache-dir .uv-cache run python -m autodesign run \
  "Create a dense academic conference poster from the attached paper." \
  --from-file /absolute/path/to/paper.pdf \
  --template cvpr-landscape

Inspect final/poster.html, final/preview.png, the final manifest, and run_events.jsonl. A file can exist after a fallback, so terminal status and validation feedback remain part of the result.

Use a visual reference 
uv --cache-dir .uv-cache run python -m autodesign run \
  "Create a paper poster using the reference's visual system." \
  --from-file /absolute/path/to/paper.pdf \
  --reference-poster /absolute/path/to/reference.png

Reference posters transfer visual systems only. Their text, claims, logos, QR codes, figures, tables, and links never become paper evidence.

One paper. Every artifact you need next.

A paper PDF flows through AutoDesign into a webpage, slides, poster, and narrated video

Finish the paper once. Paper All-in-One packages the same source into everything that usually comes next: a promotional webpage, conference slides, an academic poster, and a narrated video with timed subtitles. No need to rebuild the paper's story for each format.

Generate the complete paper suite ↗

Watch AutoDesign in action

Follow a guided local walkthrough: configure the Workbench, launch Paper All-in-One, inspect the run, and enter each editable canvas. You can also try the online demo in your browser; for the complete, most reliable experience, we recommend installing AutoDesign locally.

Guided local walkthrough · Paper All-in-One → editable canvases

https://github.com/user-attachments/assets/69c25973-fedf-4273-aa33-6bd3e409c692

Open the academic poster wall

Claude 4.8 authoring route

LongCat-Next academic poster generated through the Claude 4.8 authoring route Underwater SAM academic poster generated through the Claude 4.8 authoring route M87 Event Horizon Telescope academic poster generated through the Claude 4.8 authoring route

Codex GPT-5.5 xhigh authoring route

Denoising diffusion probabilistic models poster generated through the Codex GPT-5.5 xhigh authoring route Lung adenocarcinoma poster generated through the Codex GPT-5.5 xhigh authoring route Economic complexity poster generated through the Codex GPT-5.5 xhigh authoring route

Why AutoDesign

Method: meta-harness optimization

A design harness is the system around a fixed LLM or MLLM that turns a multimodal source into a human-facing artifact through an execution trajectory. A meta-harness improves that surrounding system. AutoDesign therefore learns from complete rollouts while keeping the underlying model weights fixed. Before autonomous optimization, an evaluator coding agent uses human-annotated reference artifacts across seven quality dimensions to implement a fixed optimization-time evaluator. It combines rule-based checks with VLM judgments and remains distinct from the frozen PosterBench protocol used for final system comparison.

Latest paper Figure 1: meta-harness optimization trajectory and DesignHarness gains across seven fixed configurations

Three stages of autonomous DesignHarness evolution followed by human-in-the-loop refinement

Autonomous outer-loop iterations evolve the harness through rollout, evaluation, one-component update proposals, and acceptance. After autonomous optimization reaches a plateau, optional Human-in-the-loop guidance can redirect the search and further improve production poster quality.

Two nested feedback loops

Loop What it improves Evidence and update
Inner loop · artifact generation One editable artifact under a fixed design harness A Designer revises the artifact; a Critic returns feedback; their interactions form an execution trajectory
Outer loop · harness optimization The reusable design harness across tasks The MetaHarnessOptimizer analyzes trajectories, evaluator scores, the persistent optimization record, and optional human guidance

Every outer-loop iteration follows four stages: rollout → evaluation → update proposal → acceptance. The optimizer acts as a planner and code editor, updates exactly one harness component, and retains the candidate only when training performance improves without reducing performance on an independent development set. Development trajectories are hidden from the update proposer.

AutoDesign meta-harness method with rollout evidence, five harness components, optimizer roles, optional human guidance, and a train-development acceptance gate

Human-in-the-loop guidance is optional. A user can give the planner observations or high-level directions to redirect a stalled search; explicit human input can also correct a systematic evaluator bias. Without guidance, the outer loop runs autonomously.

Five design-harness components

Component Elements optimized by the meta-harness
Context and Memory Multimodal source management, task prompts, skills, reusable assets, and persistent revision state
Tools and Specifications Tools and editable-artifact specifications for layout, typography, and provenance
Execution Runtime The workspace and runtime for authoring, rendering, validating, and exporting
Orchestration Task routing, attempt budgets, loop control, candidate selection, fallback, and finalization
Evaluation and Feedback Rule-based validation, model-based critique, and localized revision feedback

The optimized DesignHarness

Meta-harness optimization yields DesignHarness, the reusable artifact-producing system. Its four stages are source ingestion, iterative artifact generation and revision, validation with dual critics, and finalization. Paper metadata, claims, figures, tables, and source locations become provenance-aware context; a coding-agent Designer edits native HTML; a rule-based validator and VLM critic return localized feedback; and the best valid candidate is made self-contained for delivery.

The current implementation permits up to 12 refinement attempts. Blocking checks cover unsafe or missing assets, broken provenance, severe overflow or overlap, and required typography or layout constraints. If no candidate passes within the budget, the retained attempt history supports a constrained fallback before the same finalization stage.

DesignHarness stages from provenance-aware paper ingestion through editable generation, rule validation, VLM critique, and finalization

Five selected attempts from one AutoDesign poster trajectory, from a clipped first draft to the accepted ninth attempt

The latest paper traces one poster run through five selected attempts. The critic identifies a clipped analysis lane at A1; A3 restores the fit, A5 refits the header, A6 rescales evidence, and A9 preserves the repaired composition and is accepted. The trajectory shows that diagnostics drive localized edits while valid layout and source-derived content survive across revisions.

PosterBench leaderboard

PosterBench evaluates a 100-paper large set and a fixed 10-paper small set across AI/ML, biomedicine and health, climate and earth environment, economics and policy, and physics and astronomy. Every output is rendered to a common poster format before scoring.

The metadata-only manifests are released on Hugging Face as YaxinLuo/PosterBench and YaxinLuo/PosterBench-mini. They can be downloaded or loaded directly with datasets without redistributing the underlying paper PDFs.

The seven dimensions are Faithfulness, Coverage, Density, Visual Evidence, Layout, Readability, and Aesthetics, weighted 10/10/15/10/20/25/10. Programmatic evidence and source-conditioned VLM judgments are aggregated first; then the strictest active ceiling for severe layout damage, insufficient presentation viability, confirmed visible failure, or protected render integrity is applied to each poster.

PosterBench evaluation protocol with localized programmatic audits, seven source-conditioned dimensions, and a protected render-integrity gate

Full-Scale Benchmark Main Track · 100 papers

AutoDesign achieves the two highest PosterBench Scores. With Claude Code and Claude 4.8 fixed, it scores 78.32, exceeding Claude Design by 7.45 points and OpenDesign by 8.87 points.

AutoDesign (Yaxin9Luo/AutoDesign)

Searching Database…

245 ·GitHub
AutoDesign
Что это такое?Опенсорсный дизайн-агент и harness, который превращает один PDF научной статьи в четыре редактируемых артефакта: постер, слайды, веб-страницу и видео с озвучкой. Заявляет, что превосходит Claude Design в создании материалов для научной коммуникации.
Это веб-приложение или что-то, что работает внутри моего кодинг-агента?И то, и другое. Основной продукт — локальная веб-мастерская (бэкенд плюс интерфейс в браузере на localhost:5173); отдельные скиллы Poster, PPT, Webpage и Video также устанавливаются прямо в кодинг-агенты вроде Codex или Claude Code, без сервера.
Как его запустить?Запустите лаунчер одной командой (нужны Node.js 22+ и ffmpeg/ffprobe) — он установится в ~/.local/share/autodesign и откроет браузер; 'autodesign doctor' проверит окружение. Можно также собрать из исходников (Python 3.10+, uv, npm) и открыть localhost:5173.
Что нужно установить заранее и нужны ли API-ключи?Как минимум Node.js 22+ и ffmpeg/ffprobe; для сборки из исходников также нужны Python 3.10+, uv и npm. Да, ключи LLM-провайдера обязательны — задайте их в .env или в панели настроек веб-интерфейса.
Получу ли я готовый постер с первого раза?Не всегда — для каждого артефакта делается до 12 попыток доработки, и если ни одна не проходит валидацию, сохраняется ограниченный запасной вариант. Файл может появиться именно после такого запасного варианта, поэтому смотрите статус в терминале и отчёты валидации.
Всё остаётся на моём компьютере?Проект работает по принципу local-first: события, манифесты, кандидаты, отчёты валидации и итоговые файлы остаются на вашей машине. Есть и онлайн-демо, но в README рекомендуется локальная установка для полного функционала.
Чем это отличается от того, чтобы просто попросить Claude Design сделать постер?На бенчмарке PosterBench с фиксированными Claude Code и Claude 4.8 он набирает 78,32 — на 7,45 балла выше Claude Design и на 8,87 выше OpenDesign. Кроме того, результаты остаются редактируемым HTML, а не превращаются в один плоский файл.
Что именно получается на выходе из одной статьи?Трек Paper All-in-One упаковывает один и тот же исходник в промо-веб-страницу, слайды для конференции, академический постер и видео с озвучкой и синхронными субтитрами, причём происхождение утверждений, рисунков и таблиц сохраняется рядом с результатом.
Что такое meta-harness оптимизация?Она улучшает harness, а не веса модели: оптимизатор анализирует полные траектории, обновляет ровно один компонент harness за итерацию и оставляет изменение, только если качество на обучении растёт, не ухудшая результаты на dev-наборе.
Можно ли использовать постер прошлого года как образец стиля?Да — референс переносит только свою визуальную систему. Его текст, утверждения, логотипы, QR-коды, рисунки, таблицы и ссылки никогда не становятся доказательной базой для статьи.
What are similar projects?Proteus, Raven, Recuris, loopx, CowAgent, rrsi, EvoTrace, oil-ui
Which topics is it in?long-horizon-agents, rsi, self-evolving, vlm, design-tools, agent-harness, agentic, coding-agents, multimodal, video-generation, design, skill
Creator
Yaxin9Luo
Size
44.4 MB
Stars
245
Updated
CommitAug 25Releaseagent-skills-v0.2.0Aug 19
Age
2mo
Commits · 90 Days