convez.ai
SkillEscrita & ConteúdoAvançado
AmoryMing

Deep-Decode: Fábrica de Conteúdo Multiformat com IA

AmoryMing/deep-decode

De um URL para 5 formatos de conteúdo pronto para publicar automaticamente

Ver no GitHub
InicialPluginScore · de 0 a 100

Transforma uma ideia ou URL em uma suíte completa de conteúdo: artigo profundo, infográficos em estilo HQ, documento ilustrado, podcast e vídeo. Opera via pipeline determinístico com grafo de skills declarativo e 155+ peças reais já produzidas, publicadas em múltiplos canais.

Instalação

$npx convez-ai AmoryMing/deep-decode

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
1
Acompanhando
0
Saúde
42%
Último commit
há 4 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Agências de conteúdo gerando múltiplos formatos a partir de uma fonte única
  • Marketing de produtos com cobertura de imprensa multicanal (email, WeChat, redes sociais)
  • Pesquisadores criando documentação visual e áudio complementar de tópicos complexos

Veredito

Projeto impressionante de orquestração de conteúdo com rigor de contrato e operação determinística. Solução robusta para equipes que precisam escalar produção multiformat sem perder consistência; exige curva de aprendizado em grafo de skills e infraestrutura.

README

A tradução para o português deste README ainda está na fila.

Deep-Decode · AI Content Factory

One idea or URL in. A full multi-format content suite out.

Deep-dive article · comic-style infographics · illustrated doc · podcast · video — produced by a deterministic skill-graph pipeline, then driven to publish-ready on every channel.

Stars Last commit Decoded pieces Built with Claude Code Next.js License

English · 简体中文


What is this

Deep-Decode is an opinionated, end-to-end AI content factory.

Give it a blog post, a tweet, a product, a person, or a trending topic. It researches the source, writes an opinion-dense deep-dive (not a translation, not a summary), draws comic-style infographics, lays out an illustrated document, voices a podcast, cuts a video, and pushes every format to publish-ready across email, WeChat, Xiaohongshu, Video Accounts and Douyin.

The whole run is orchestrated by a declarative skill graph and a deterministic runner that refuses to advance until each step's artifact passes its contract. No silently-skipped steps. File = state. Failure is visible.

This is not a demo. The output/ folder holds 155+ real pieces shipped through the pipeline.

Why it's different from "ask an LLM to write a post"

Generic LLM promptDeep-Decode
StanceTranslates / summarizesReconstructs the argument with its own framework; opinion density > information density
RigorSingle source, no checkingCross-checks 2–3 external signals; coins a name when it spots an unnamed phenomenon
OutputOne blob of text5 formats from one source of truth — article, infographics, doc, podcast, video
Reliability"Hope the agent finished"Declarative graph + hard artifact gates; a missing file stops the run
ReachCopy-paste yourselfAuto-driven to draft-ready on 5 Chinese platforms

Product tour

Every figure below is generated by the GPT-Image (img2) backend — accurate Chinese typography, a consistent palette, magazine-grade layout. These are real, recent pieces (June 2026), straight from the pipeline. Click any image to read the decode.

Each deep run emits a full img2 figure set (cover + per-section illustrations, like the hero grid above), plus a podcast (podcast.mp3), a video (video.mp4), an illustrated Word doc (.docx), and magazine-style Xiaohongshu card decks — all addressed by the same output/<slug>/ contract. The backend is switchable per project via image_backend: gpt-image.

How it works — the skill graph

The old pipeline was ~18 prose steps in a doc. Agents running a deep chain would silently get lost — skip a step, use the wrong template, stop at an email draft instead of shipping. So the flow became a declarative graph (skillgraph.yaml) traversed by a deterministic runner (tools/pipeline.py).

            ┌──────────────────────────────────────────────────────────┐
            │  skillgraph.yaml  —  declarative 3-layer graph            │
            │                                                          │
   compounds│  decode · brief · practice · distribute-all   ← you drive │
            │      ▲  (pick a playbook + confirm Strategy Spec)         │
  molecules │  article · infographic set · podcast · video · doc        │
            │      ▲  (each = one deliverable)                          │
     atoms  │  svg→png · tts · imagegen · tone-lint · send              │
            │      ▲  (single deterministic action)                     │
            └──────────────────────────────────────────────────────────┘
                         │  topological order computed by
                         ▼
            tools/pipeline.py  —  the ONLY authority on step order
                 status · next · gate · verify
  • Three layers. atoms (one deterministic action) → molecules (one deliverable) → compounds (a full playbook a human drives).
  • You don't count steps. The runner computes the topological order from depends_on edges.
  • Done = artifact exists and passes its contract. Contracts include file_exists, min_bytes, png_for_each_svg, audio_visual_sync, tone_match (banned-word + voice lint), and channel_draft_ready. A missing artifact = stuck; the runner won't continue.
  • Humans steer at the compound layer only — choose the playbook, confirm the Strategy Spec. Everything else the runner pulls along.
cd output/2026-06-08-some-slug
python3 ../../tools/pipeline.py status   # whole graph + ✓/✗ + next step
python3 ../../tools/pipeline.py next      # "what node runs now?"
python3 ../../tools/pipeline.py gate m.article   # verify the artifact contract

Examples

A few of the 155+ pieces (titles translated; sources are decoded in Chinese):

PieceSource decoded
The referee can't be the player: why Claude builds itself a harnessClaude Code blog
The bottleneck moved: when writing code is free, re-sequence the orgEngineering essay
Fei-Fei Li sets the rules for "world models": not text-to-video, a POMDP loopFei-Fei Li
Ideogram v4.0: when text-to-image learns to spell and follow structureLaunch + benchmarks
Zhipu returns to A-shares and renames Z.AI: China's second capital pathFiling + reporting
Karpathy joins Anthropic: star-individual migration as a roadmap signalTechCrunch
0.2 points and a 7× price gap — DeepSeek V4 compresses the paradigm war into one math problemOfficial + 3rd-party benchmarks

Browse them all in output/, or through the web workbench below.

The web workbench

A Next.js 15 app (web/) turns the factory into a three-in-one site:

SectionRouteAccessWhat
Portal/, /post/[slug]publicevery decoded piece — article + infographics + podcast + video
Process/processpublicthe three verbs + the live skill-graph (reads the real skillgraph.yaml)
Admin/adminloginschedule / in-progress / queue / revenue dashboard

Static-generated, zero media in the deploy bundle — images/audio are rewritten to a jsDelivr CDN backed by this repo, video to GitHub raw. Deploys to Vercel with deploy as the production branch.

cd web
npm install
cp .env.example .env.local
npm run dev          # http://localhost:3000

Repository layout

deep-decode/
├── .claude/skills/   # the skills: deep-decode, polish, visual, podcast, video, distribute…
├── skillgraph.yaml   # the declarative 3-layer graph (single source of step order)
├── tools/            # pipeline.py runner + atoms (tone_lint, tts_atom, imagegen_relay…)
├── output/           # 155+ shipped pieces — one folder each (article + media)
├── wiki/             # structured knowledge: topics / sources / concepts / published
├── schedule/         # queue · in-progress · published · calendar
├── styles/           # voice + feedback + best-of, multiple kits
├── readers/          # audience personas (drives tone)
├── templates/        # content + project templates
└── web/              # Next.js 15 portal + process view + admin

Quick start

Deep-Decode runs as a Claude Code skill system.

  1. Clone the repo and open it in Claude Code — the skills live under .claude/skills/.
  2. Drop source material in raw/, or hand it a URL.
  3. Say "decode this" (/deep-decode <url>) and confirm the Strategy Spec; the runner takes it from there to publish-ready.
  4. Run the web workbench (cd web && npm run dev) to browse what you've made.

Heads-up: distribution targets Chinese platforms (WeChat / Xiaohongshu / Video Accounts / Douyin) and Chinese TTS voices. The architecture — declarative skill graph, contract gates, one-source-to-many-formats — is fully general; fork the pattern for any language or channel set.

Status

Actively used in production for a Chinese AI-commentary publication. deploy is the content branch (the default branch on GitHub); see web/README.md for why content lives on an orphan branch and how it's pushed.

License

Apache-2.0 © 2026 AmoryMing. The pipeline, skills, and tooling are free to use and adapt. The decoded articles and generated media under output/ are illustrative examples — please don't republish them as-is.


Ver o original no GitHub

Bom para

Relacionados