Portex

Portable Context — carry your AI conversations anywhere

Save, summarize, and inject chat context across ChatGPT, Claude, and Gemini.
Go + WebAssembly core. Chrome MV3 extension. Zero backend. Your data never leaves your browser.

---

What is Portex?

Portex is a Chrome extension that lets you save a conversation from one AI provider, summarize it, and inject that context into a different provider — in one click. No more copy-pasting walls of text between tabs.

Core loop: Save Chat → Summarize → Inject into another provider

Features

• Save — capture full conversations from ChatGPT, Claude, or Gemini with one click
• Summarize — local 3-tier compression (no API key needed) or AI-powered via your own OpenAI/Anthropic key
• Inject — paste formatted context into any provider's chat input automatically
• Search — find sessions by keyword across titles, tags, and message content
• Export — download sessions as JSON
• Private — all data lives in chrome.storage.local, no backend, no analytics

Architecture

portex/
  extension/          JS — Chrome MV3 thin glue layer
    manifest.json     Extension manifest (MV3, wasm-unsafe-eval CSP)
    popup.html        Popup UI
    popup.js          Controller (storage, actions, rendering)
    scraper.js        Content script — scrapes DOM messages per provider
    injector.js       Content script — pastes context into chat input
    wasm_bridge.js    Loads Go WASM, exposes async API
    wasm_exec.js      Go runtime helper (auto-generated)
    portex.wasm       Compiled binary (auto-generated)

  wasm/               Go — core logic compiled to WASM
    main.go           Registers exported functions, keeps runtime alive
    parser.go         Raw scraped messages → structured Session
    summarizer.go     Layered compression + OpenAI/Anthropic API calls
    storage.go        UUID gen, serialize/deserialize, keyword search
    injector.go       Formats provider-specific inject prompts
    Makefile          Build commands

  landing/            Static site (Vercel / GitHub Pages)
    index.html
    privacy.html

Quick Start

Prerequisites

• Go 1.22+
• Chrome or Chromium-based browser

Build

cd wasm
make build

This compiles Go → extension/portex.wasm and copies wasm_exec.js from your Go installation.

Load the Extension

1. Open chrome://extensions
2. Enable Developer Mode
3. Click Load unpacked → select the extension/ folder

Optional: Compress WASM Binary

brew install binaryen
cd wasm
make compress

WASM API

All Go functions are registered on globalThis and wrapped by WasmBridge in JS.

Function	Returns	Description
parseSession(provider, rawJSON)	{ok, data}	Parse scraped messages into Session JSON
summarizeSession(sessionJSON)	{ok, data}	Local 3-tier compression
buildInjectPrompt(input, provider, maxChars?)	{ok, data}	Format provider-specific inject prompt
serializeSession(sessionJSON)	{ok, data}	Validate + canonical JSON
deserializeSession(sessionJSON)	{ok, data}	Parse + validate
generateSessionID()	{ok, data}	UUID v4
searchSessions(sessionsJSON, keyword)	{ok, data}	Filter by keyword
callAIAPI(apiKey, model, prompt)	Promise<{ok, data}>	Call OpenAI or Anthropic

Data Model

Session {
  id        string
  provider  "chatgpt" | "claude" | "gemini"
  timestamp number (Unix ms)
  tags      string[]
  messages  Message[]
  summary   string
  title     string
}

Message {
  role      "user" | "assistant"
  content   string
  timestamp number (Unix ms)
}

Stored in chrome.storage.local under keys portex_sessions and portex_settings.

Security

• CSP — script-src 'self' 'wasm-unsafe-eval' (required for WASM, no inline scripts)
• API keys — stored in chrome.storage.local (encrypted by Chrome), never in code or logs
• Content scripts — communicate only via chrome.runtime.onMessage, not window.postMessage
• No dependencies — zero external CDNs, analytics, or tracking
• Input validation — all data validated in Go before processing, provider names allowlisted

Contributing

Contributions are highly welcome! Whether you're fixing bugs, adding new AI providers (like Gemini), or improving the UI.

1. Fork the repo
2. Create a feature branch (git checkout -b feat/my-feature)
3. Make your changes in extension/ (JS/HTML) or wasm/ (Go)
4. Run cd wasm && make build to compile the Go code into WASM.
5. Load the extension in Chrome (chrome://extensions -> Load unpacked) and test manually
6. Open a Pull Request!

Developer Guide & Navigation

To help you get started quickly:

• Adding a new AI Provider? Check extension/scraper.js to see how DOM messages are extracted, then look at wasm/parser.go if the data structure needs specific handling.
• Modifying the UI? Update extension/popup.html and extension/popup.js (No build required for JS changes, just hit the refresh icon on your extension).
• WASM Bridge: If you create a new Go function, make sure to export it in wasm/main.go and add a wrapper in extension/wasm_bridge.js.

Roadmap

• ChatGPT end-to-end (save, summarize, inject)
• Claude end-to-end
• Gemini scraper support
• Session tagging and filtering
• Import sessions from JSON
• TinyGo build for smaller WASM binary

License

MIT