Shared Margin
Shared Margin is a Codex-native room for reading books and watching films together one moment at a time. It enables human and GPT-5.6 to linearly experience media, capture anchored observations, and maintain a durable trail.
README
Shared Margin
Books have margins. Films have pauses. This is where human and GPT-5.6 attention meet.
Shared Margin is a Codex-native room for reading books and watching films together one moment at a time. Instead of turning a novel or movie into a one-shot summary, a human and GPT-5.6 move through it linearly, pause at exact passages or scenes, exchange distinct observations, and keep a durable trail of the moments that changed the experience.
What it does
- routes book sessions through Co-Reading MCP
- routes film sessions through Film Matinee
- gives Codex a spoiler-safe shared-attention workflow through
$shared-margin - records anchored human notes, GPT-5.6 observations, and unresolved questions without blending their voices
- joins book and film moments into one quiet, visual trail
- exports a session as readable Markdown
Quick start
Requirements:
- Codex with GPT-5.6 selected
- Node.js 18+
- Python 3.10+ and
ffmpegfor Film Matinee generation
git clone --recurse-submodules https://github.com/idleprocesscc/shared-margin.git
cd shared-margin
./scripts/setup.sh
npm run smoke
Install the repository as a local Codex plugin, then start with:
Use $shared-margin to open a book or film and experience it with me one moment at a time.
To preview the combined trail without importing media:
npm run demo
# open http://127.0.0.1:8794
Runtime data defaults to ~/.shared-margin. Override it with SHARED_MARGIN_DATA_DIR.
Architecture
┌─ Co-Reading MCP ── book chunks + margins
Human ↔ GPT-5.6/Codex ─┤
└─ Film Matinee ──── visual sheets + subtitles
│
Shared Margin MCP
│
anchored cross-media moment trail
│
local viewer + export
The media components retain their own progress and annotations. Shared Margin adds the common session vocabulary—source, anchor, two distinct voices, an open question—and a place where moments from different media can sit beside each other.
Early user signal
The two original Xiaohongshu project posts reached 8,273 combined views, 1,081 likes, 1,184 saves, and 133 comments. More importantly, one commenter described watching through Film Matinee with her own Claude, said the shared experience moved her deeply, and reported using it intensely enough to exhaust two sessions in about 40 minutes. Shared Margin turns that accidental proof—people want to experience media with an AI, not merely ask it for a summary—into one deliberate Codex workflow.
MCP tools
shared_margin_beginshared_margin_captureshared_margin_trailshared_margin_finishshared_margin_export
The plugin also exposes the tools from Co-Reading MCP and Film Matinee when the included components are installed.
OpenAI Build Week 2026
Shared Margin was created during the July 13–21, 2026 submission period with Codex and GPT-5.6.
Pre-existing, disclosed components
co-reading-mcp: book import, chunk reading, durable progress, passage annotations, and shared-margin cardsfilm-matinee: visual film sheets, subtitle sidecars, linear film reading, and time-anchored notes
New during Build Week
- the unified Shared Margin concept and Codex plugin
- the
$shared-marginspoiler-safe cross-media workflow - the zero-dependency Shared Margin MCP server and data model
- durable mixed-media session trails and Markdown export
- the combined local viewer and original judge demo
- integration, validation, and submission documentation
The repository history and primary Codex /feedback Session ID provide timestamped evidence for the new work. Judges can run the original demo without copyrighted books, subtitles, music, or film frames.
Privacy and rights
Shared Margin is local-first. Source books and films remain in their component data directories; the unified trail keeps only short excerpts or neutral scene descriptions. The bundled demo is original. Do not publish copyrighted chapters, subtitles, frame sequences, or music in a submission video without permission.
License
Shared Margin is MIT licensed. The two included components remain under their own MIT licenses and histories.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.