auto-company
An experimental multi-agent system where a team of AI agents runs an autonomous company, handling product ideas, coding, deployment and marketing with no human involved. Built on Claude Code Agent Teams, still in testing and macOS only.
Copies AI chat conversations from ChatGPT, Claude, Gemini and similar tools as clean Markdown context bundles, with one-click copying from the sidebar and all processing done locally.
Claude Code
$ npx -y skills add nicepkg/ctxport --agent claude-code
Codex
$ npx skills add nicepkg/ctxport -a codex
Cursor
$ npx skills add nicepkg/ctxport -a cursor
Gemini CLI
$ npx skills add nicepkg/ctxport -a gemini-cli
GitHub Copilot
$ npx skills add nicepkg/ctxport -a github-copilot
OpenCode
$ npx skills add nicepkg/ctxport -a opencode
Windsurf
$ npx skills add nicepkg/ctxport -a windsurf
Your AI conversations deserve a better clipboard.
简体中文 | English
Supported Platforms
Get Started · Features · Documentation
You just spent 45 minutes in a deep conversation with ChatGPT. You found the perfect architecture for your project. Now you need to share it with Claude for implementation.
What do you do?
Ctrl+A, Ctrl+C? You get a mess of HTML artifacts, broken formatting, and missing code blocks.
Copy each message manually? Life is too short.
Screenshot it? Text in images is where knowledge goes to die.
AI conversations are the new unit of knowledge work. But moving them between tools feels like faxing a PDF in 2026.
CtxPort fixes this. One click, and your entire conversation becomes a clean, structured Markdown document -- ready to paste into any AI tool, editor, or knowledge base.
| Without CtxPort | With CtxPort | |
|---|---|---|
| Copy a conversation | Select all, copy, paste, fix formatting for 10 min | One click. Done. |
| Move context between AIs | Re-type your whole conversation history | Paste the Context Bundle, keep going |
| Save a conversation | Bookmark it and pray the URL survives | Structured Markdown you own forever |
| Share with your team | "Let me screenshot this..." | Share a clean .md file |
| Extract just the code | Scroll through 50 messages hunting for snippets | Code Only mode, one click |
No account required Works offline
Zero data uploaded 100% local processing
Minimal permissions Open source (MIT)
graph LR
A["AI Conversation<br/>(ChatGPT, Claude, etc.)"] --> B["CtxPort<br/>Browser Extension"]
B --> C["Context Bundle<br/>Structured Markdown"]
C --> D["Another AI Tool"]
C --> E["Your Notes / Docs"]
C --> F["Team Chat / PR"]
Alt+Shift+C)That's it. No configuration. No sign-up. No cloud.
| Feature | Description |
|---|---|
| In-Chat Copy Button | A copy button appears right in the conversation -- click to copy the entire thread |
| Sidebar List Copy | Hover over any conversation in the sidebar and copy without even opening it |
| Keyboard Shortcut | Alt+Shift+C copies the current conversation instantly |
| Multiple Formats | Full, User Only, Code Only, Compact -- pick what you need |
Most copy tools make you open a conversation first. CtxPort lets you hover over conversations in the sidebar and copy them directly. Need to grab 5 conversations for a project brief? Hover, click, hover, click. No page loads. No waiting.
Every copy produces a structured Markdown document with frontmatter metadata:
---
ctxport: v2
source: chatgpt
url: https://chatgpt.com/c/abc123
title: "Discuss REST API Authentication"
date: 2026-02-07T14:30:00Z
nodes: 24
format: full
---
## User
I'm building a SaaS product and need to choose between
API key auth and OAuth2. What do you recommend?
## Assistant
Based on your scenario, I'd recommend a layered approach...
The frontmatter tells any receiving tool exactly where this conversation came from, when it happened, and how many messages it contains. Structured context, not just raw text.
| Format | What You Get | Best For |
|---|---|---|
| Full | Complete conversation with all messages | Context transfer between AI tools |
| User Only | Only your messages (prompts) | Reusing your prompts in a different AI |
| Code Only | Extracted code blocks with language tags | Grabbing implementation snippets |
| Compact | Condensed single-paragraph messages | Quick sharing in chat or email |
Coming soon -- currently in development. Star the repo to get notified!
# Clone the repo
git clone https://github.com/nicepkg/ctxport.git
cd ctxport
# Install dependencies
pnpm install
# Build all packages
pnpm build
# Start the extension in dev mode
pnpm dev:ext
Then load the unpacked extension from apps/browser-extension/dist/chrome-mv3-dev in chrome://extensions.
Alt+Shift+CFor sidebar list copy: hover over any conversation in the left sidebar to reveal the copy icon.
ctxport/
packages/
core-schema/ # Zod schemas for Context Bundle format
core-plugins/ # Platform adapters (ChatGPT, Claude, etc.)
core-markdown/ # Markdown serialization engine
shared-ui/ # Shared React components
apps/
browser-extension/ # WXT + React 19 + Tailwind CSS 4
Built as a monorepo with pnpm workspaces and Turborepo. Each platform adapter is a self-contained plugin, making it straightforward to add new platforms.
Contributions are welcome! Whether it's a bug report, feature request, or pull request -- every contribution helps.
# Fork and clone the repo
git clone https://github.com/YOUR_USERNAME/ctxport.git
# Install dependencies
pnpm install
# Start development
pnpm dev:ext
See CONTRIBUTING.md for detailed guidelines.
MIT -- Use it however you want.
If CtxPort saves you time, consider giving it a star.
It helps others discover the project and keeps development going.
An experimental multi-agent system where a team of AI agents runs an autonomous company, handling product ideas, coding, deployment and marketing with no human involved. Built on Claude Code Agent Teams, still in testing and macOS only.
A skill that shapes Claude Code's output for readers with ADHD: it leads with the next action, numbers multi-step work, restates state across turns, drops tangents, gives specific time estimates, and makes progress visible. Invoke it with /i-h.
Gives your AI agent working knowledge of Obsidian, so it can read and write notes, build Bases database views and JSON Canvas mind maps, and turn web pages into clean Markdown for your vault.
Gives product managers step-by-step help with discovery, prioritization, strategy, launch and growth work, so they get structured outputs instead of loose text. Also includes skills for reviewing AI-built code before shipping.
Formerly: Colleague Skill / colleague-skill.
Install it with the command on its page. It includes team. Its skills do not fire on their own yet. Free and open source.
Is this plugin yours?
Claim it with GitHubSubmit a pluginPromote it