Compound Writing: Iterative AI Drafting via Interviews and Style Files

X postBrandon GellResource

Brandon Gell highlights 'Compound Writing', an approach showcased by media outlet Every. Instead of generating complete articles in one go, the model interviews the author, drafts section by section, and stores stylistic learnings in persistent files. The workflow has been released as an open-source plugin for Claude Code and Codex.

Key points

  1. Critique of the vending-machine approach: Single-prompt drafting often yields generic, robotic prose.
  2. Interview-led workflow: Prior to drafting, the model poses 15 to 30 targeted questions to extract the author's ideas.
  3. Step-by-step drafting: Outlining, drafting, and reviewing take place iteratively, one section at a time.
  4. Compounding context: Insights from drafting sessions are retained in dedicated voice and style files for future pieces.
  5. Open-source plugin: Built for Claude Code and Codex, including specialist review personas such as /hitchcock and /sorkin.

Why it matters

Maintaining a distinct personal or brand voice remains a primary hurdle in AI-assisted writing. Compound Writing shifts the focus from one-shot text generation to upfront inquiry and cumulative context building, ensuring better output consistency over time.

Analysis

This method borrows foundational software development patterns—such as iterative review cycles and file-based state management—and applies them to editorial workflows. By accumulating style rules in persistent files, it addresses the context-loss problem common to standard chat interfaces.

Original post

Brandon Gell

@bran_don_gell · 5 September 2026

If you write for a living you need to use Compound Writing. Compound Engineering is used by tens of thousands of top engineers. Pretty sure CW will be the same for writers. t.co/Z20BNjHBN5

235 likes5 replies495 bookmarks48.7k views

View on X
Selected replies (5)
  • @drakederen @bran_don_gell Without a doubt the best writing I’ve gotten with AI and probably the best plugin I’ve use all year 🙏
  • @ger_tzedek @bran_don_gell Looks super interesting, but this is a an awful, scammy, growth hacky site, that forced me to sign in to read the article… and then paywalled me.
  • @JakeBlockchain @bran_don_gell Will try it this week
  • @sabshrtali @bran_don_gell I am building t.co/fpluSo3YsC for education Soon it ll be in the list of compound named products loo 😉
  • @keskinonur @bran_don_gell Ref: t.co/FnRQIICBLe

Mentioned links

Links and tools from this entry

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Related entries

Similar in content, found by the AI search.

  • Repository:EveryInc/compound-engineering-plugin

    Compound Engineering: 33-Skill Plugin for AI Coding Agents

    Compound Engineering is an open-source plugin for AI coding agents designed to structure development workflows and systematically preserve learned context. It runs on 14 platforms, including Claude Code, Cursor, and Codex.

    24.9kstarsTypeScript

    AI & LLMs· Tool

  • Article:Peter Yang

    Open-Source Skill to Strip AI Slop and the 25/50/25 Rule for Writing

    Peter Yang has open-sourced an AI skill called /no-ai-slop for harnesses like Claude Code and Codex. The tool detects and removes over 20 repetitive patterns commonly produced by generative models. Alongside the release, Yang outlines a 25/50/25 writing process to retain editorial control.

    AI & LLMs· Tool

  • X post:Daniel

    Automating AI UGC Production with Opus 5.5 and Claude Code

    Daniel (@danclipping) outlines a workflow for building an automated AI User-Generated Content (UGC) pipeline using Opus 5.5 and Claude Code. The model functions as an orchestration layer and director handling research, scripts, prompts, and visual evaluation, while external tools manage video rendering and distribution.

    1.6kbookmarks209.2kviews

    AI & LLMs· Tutorial

  • Repository:petergyang/no-ai-slop

    no-ai-slop: Agent Skill to Strip Formulaic AI Phrasing

    The open-source project no-ai-slop detects and removes over 20 stereotypical patterns of AI writing without flattening the author's individual voice.

    7kstarsPython

    AI & LLMs· Tool

  • Repository:chaseai-yt/claudex-loop

    claudex-loop: Adversarial Plan Hardening for Claude Code and OpenAI Codex

    claudex-loop is a Claude Code skill and plugin that hardens development plans before any code is written. By placing Claude and OpenAI Codex into an adversarial cycle, competing models inspect and challenge specifications to prevent architectural flaws.

    1.6kstars

    AI & LLMs· Tool

Back to the knowledge base

Let's talk about your project

Get in touch

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.