Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Claude: Vercel Commerce Sandboxes
Learn how Claude operates in isolated Vercel sandboxes to manage a live commerce system. Discover Git-based mutations, typed commerce SDK calls, and Just Bash execution for a robust AI architecture.
This talk walks through how we run Claude inside per-session Vercel Sandboxes to operate on a live Next.js commerce system.
Each session provisions an isolated sandbox where Claude connects via the Claude Agent SDK over WebSockets. The initial prompt fetches a Git repository of the storefront, which becomes the working workspace. From there, Claude proposes structured mutations that are applied incrementally as diffs and commits, with optional deployment to Vercel.
Commerce state (products, pricing, checkout) is never modified via arbitrary code edits. Instead, mutations flow through typed calls exposed by our Commerce SDK. UI changes can be generative; transactional operations must pass validated primitives.
We’re evolving the system toward an in-memory project file representation to reduce cold starts and avoid long sandbox wakeups. Mutations are first applied to this in-memory graph, then executed via Just Bash inside the sandbox, enabling background transitions while keeping the live instance responsive.
The demo focuses on the WebSocket orchestration loop, sandbox isolation boundaries, tool schema design, and the failure cases we encountered while ensuring the commerce system remains consistent and deployable at all times.
AI-native Next.js commerce template featuring Stripe and typesafe Commerce SDK.
- Vercel SandboxIsolated execution environments for running untrusted code and AI-generated scripts with zero-config infrastructure.Vercel Sandbox delivers secure, ephemeral micro-VMs designed to execute dynamic code (Node.js, Python, and Bash) generated by LLMs. It pairs with the Vercel AI SDK to enable safe tool-calling: models can run complex data visualizations or math operations without compromising the host environment. Each instance boots in under 100ms and scales automatically (handling the heavy lifting of infrastructure management so teams focus on logic). This setup provides the isolation needed for production AI agents to interact with real-world data securely.
- Claude Agent SDKThe Claude Agent SDK (Python, TypeScript) transforms Claude into an autonomous, production-ready agent by granting it controlled access to a computer: file operations, Bash execution, and external tool integration via the Model Context Protocol (MCP).This SDK provides the core agent harness that powers Claude Code, enabling developers to build sophisticated, general-purpose agents. It offers official TypeScript and Python libraries for seamless integration. Agents gain powerful, safe capabilities: reading and writing files, executing sandboxed Bash commands, and leveraging a rich tool ecosystem. The framework includes essential production features (e.g., automatic context management, granular permission modes like `acceptEdits`, and built-in cost tracking). Use the open Model Context Protocol (MCP) to connect custom tools, databases, and APIs, extending Claude’s utility far beyond simple chat to complex, multi-step workflows.
- WebSocketsSecures a persistent, full-duplex TCP connection via a single HTTP handshake, delivering low-latency, bidirectional data streaming for real-time applications.WebSockets establishes a continuous, two-way communication channel: a critical shift from the request/response cycle of HTTP/1.1. The connection initiates with an HTTP `Upgrade` handshake, switching the protocol to `wss` (secure) on port 443. This persistent link minimizes network overhead, eliminating the need for inefficient HTTP polling. The protocol, standardized as IETF RFC 6455 in 2011, ensures near-instant data transfer: vital for high-performance use cases (e.g., live chat, collaborative editing, financial market data).
- gitGit is the distributed version control system (DVCS) that tracks source code changes, ensuring data integrity and enabling non-linear development workflows.Git is the free, open-source distributed version control system (DVCS) created by Linus Torvalds in 2005 to manage the Linux kernel. Engineered for speed and efficiency, it handles projects from small to extremely large, storing the entire 1.4 million commit history of the Linux project in only 5.5 GB . Its core design supports non-linear development (branching/merging) and guarantees data integrity via cryptographic hashing. According to a 2022 Stack Overflow survey, 96% of professional developers use Git, making it the industry standard for collaborative software development .
- Just BashA modular framework and standard library for Bash 4.4+ that introduces structured imports and namespaced functions.Just Bash turns standard shell scripts into robust applications by providing a modular architecture for Bash 4.4 and higher. It replaces messy sourcing with a clean import system (using the `import` command) and includes 50+ optimized modules for tasks like JSON parsing, string manipulation, and logging. The framework ensures reliability through strict error trapping and namespaced variables: preventing the global scope pollution common in traditional scripts. It is a zero-dependency solution (requiring only the shell itself) designed for high-performance automation where speed and maintainability are non-negotiable.
Related projects
Jules: An AI Agent That Thinks, Argues, and Writes
Austin
See how an AI agent can think, argue, and write with you. This talk demos structured reasoning, persistent…
Claude in the Channel: Building an Autonomous Frontend Agent for No-Code Teams
Tokyo
Learn how to build an autonomous frontend agent using Slack and Claude to enable designers to ship production…
What Claude Code Should Be
San Diego
rm -rf salesforce — Claude Runs Deal Flow Now
Seattle
Learn how to replace your CRM with files and Claude agents. This talk demonstrates a git-backed knowledge base…
Watch a Sales Guy Ship Agentic Infrastructure
Toronto
See how an MCP server uses a merchant's disclose.json to provide AI agents with verifiable operational data like…
Multi-agent Claude Code at scale: building Parthas at Intercom
Dublin
See a live demo of Parthas, a multi-agent orchestrator for Claude Code, coordinating parallel coding agents through software…