A simple, secure API for storing and retrieving AI conversation contexts using Cloudflare Workers and R2 storage.

CloudContext solves a basic problem: AI assistants forget everything between conversations. This service provides a persistent storage layer for AI contexts, conversation history, and user preferences that survives across sessions.

I built this because I was tired of having to re-explain context to AI assistants every time I started a new conversation. It runs on Cloudflare's infrastructure, so it's fast and has good global coverage.

• Store AI contexts, conversation history, and user data via a REST API
• Data is stored in Cloudflare R2 (S3-compatible object storage)
• Basic authentication with API keys
• Versioning support for tracking context changes
• JavaScript/TypeScript client library included

• REST API for storing/retrieving contexts
• Bearer token authentication
• Context versioning and history
• CORS support for browser usage
• JavaScript client library
• Automated deployment script

• Cloudflare account
• Wrangler CLI installed
• Basic knowledge of Cloudflare Workers

1. Clone this repository
2. Copy wrangler.example.toml to wrangler.toml and configure your settings
3. Run the setup script:

./setup.sh

Or deploy manually:

npm install -g wrangler
wrangler deploy

import CloudContext from './clients/javascript/index.js';

const client = new CloudContext({
  baseUrl: 'https://your-worker.your-subdomain.workers.dev',
  apiKey: 'your-api-key'
});

// Save some context
await client.save({
  conversation: ['Hello', 'Hi there!'],
  preferences: { theme: 'dark' }
});

// Retrieve it later
const context = await client.get();
console.log(context.preferences.theme); // 'dark'

• POST /api/context - Save context data
• GET /api/context - Retrieve context data
• GET /api/context/list - List all contexts for a user
• GET /api/context/version - Get context version history
• POST /api/context/restore - Restore a previous version
• POST /api/context/sync - Sync context data
• GET /api/health - Health check

This is a working implementation but still evolving. The core functionality is stable, but I'm continuing to add features and improve the API.

Currently implemented:

• ✅ Core storage and retrieval
• ✅ Authentication system
• ✅ Context versioning
• ✅ JavaScript client
• ✅ Automated deployment

• JavaScript/TypeScript: Ready to use (see clients/javascript/)
• Python: Planned
• Go: Planned

Copy wrangler.example.toml to wrangler.toml and update with your settings:

name = "your-cloudcontext-worker"
main = "src/worker.js"

[env.production]
vars = { ENVIRONMENT = "production" }

[[env.production.r2_buckets]]
binding = "CONTEXT_BUCKET"
bucket_name = "your-context-bucket"

# Start local development server
wrangler dev

# Run tests
npm test

# Deploy to production
wrangler deploy

Using Cloudflare's free tier, this should handle thousands of requests per day at no cost. R2 storage is very affordable - typically under $1/month for most personal use cases.

Pull requests welcome. This is a fairly simple project, so please keep contributions focused and well-tested.

MIT License - see LICENSE file.