# Claude Code SDK Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/claude-code-sdk
> Markdown URL: https://aitinkerers.org/technologies/claude-code-sdk.md
> Technology record last updated: 2026-02-26T05:57:40Z
> Generated: 2026-09-20T16:47:06Z

The programmatic interface for Claude Code: build production-ready AI agents with full access to code execution and rich tool use.

This is the Claude Code SDK (also known as the Claude Agent SDK): the core agent harness now available as a library (TypeScript and Python) for production systems. It lets you integrate Claude's agentic capabilities—like file operations, code execution, and web search—directly into your developer tools and scripts. Leverage built-in features like automatic context management and the Model Context Protocol (MCP) for powerful, enterprise-ready integrations with external systems like Jira or Slack. Deploy sophisticated agents, such as SRE agents for diagnosing production issues or security review bots for auditing code, with fine-grained control over permissions and capabilities.

- Official technology site: https://docs.anthropic.com/en/docs/claude-code/sdk
- Public AI Tinkerers demos and talks: 7
- Result page: 1 of 1

## Recent Public Talks and Demos

### [Live fixing an agent defect](https://paris.aitinkerers.org/talks/rsvp_V8gBM8cPPv4)

Ashley is my personal assistant for the AI Tinkerers community. Today she unsubscribed our Paris organizer, Abdoulaye, from ten different things! She should definitely not have done that! This will be a nice thing to show the community since I'm here with Abdoulaye in Paris now.

- Event context: Offstage w/ Docker feat ClickHouse — 2026-07-07 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_V8gBM8cPPv4

### [Playwright MCP vs CLI? Benchmarking for performance & effectiveness](https://denver-boulder.aitinkerers.org/talks/rsvp_ZkszEZey4aY)

I benchmarked an agent using Playwright MCP vs Playwright CLI vs agent-browser. The CLIs were much more context efficient, but they were--surprisingly--also much slower! I dug into why that might be and the answer came down to tool calls, which turned out to be driven directly from the interface designs of each tool. The MCP used up a lot of the context window by returning all state information on every run. The CLI and agent-browser both returned much less information, leaving it up to the agent to directly query the information it needed. That meant less of the context window was used, but more tool calls were required and tool calls were the most direct corollary to timing. The biggest surprises were that context efficient != faster, and it _also_ didn't mean cheaper. ![](https://outpost.ranger.net/_astro/2-runtime-dots.RG7TW8g3_1jz1a2.webp) ![](https://outpost.ranger.net/_astro/6-tool-call-type-bars.CD3GEom5_Z2lR38P.webp) ![](https://outpost.ranger.net/_astro/9-cost-slope.CQX0e7Dr_1twtqz.webp)

- Event context: AI Tinkerers Denver Meetup - March 25, 2026 — 2026-03-25 — Denver
- Public talk page: https://denver-boulder.aitinkerers.org/talks/rsvp_ZkszEZey4aY

### [The IDE is Dead. Vulcan Ships Our Code Now.](https://london.aitinkerers.org/talks/rsvp_lcPJx56YCAU)

Vulcan is an autonomous AI dev system we built at a stealth startup. It orchestrates Claude Code agents on Mac Minis to execute parallelizable development plans, create PRs, and autonomously fix its own CI failures. We haven't opened an IDE in months. I'll walk through: the plan/phase architecture with dependency graphs for concurrent Claude instances, the worker daemon that spawns and manages Claude Code subprocesses, the self-healing CI loop (including BFS-based loop detection to stop infinite fixes), and how MCP tools let Claude manage its own PR lifecycle through our Django backend. Live demo: submitting a plan, workers claiming phases, CI breaking, and Vulcan fixing itself.

- Event context: AI Tinkerers London - 2nd March featuring AMA with Tom Occhino, CPO at Vercel — 2026-03-02 — London
- Public talk page: https://london.aitinkerers.org/talks/rsvp_lcPJx56YCAU

### [AI + Filesystem: Building Claude-powered File Manipulation Agents](https://montreal.aitinkerers.org/talks/rsvp_-GRVa1txV6o)

In this talk, I’ll show how the Claude Code SDK gives agents real file‑system power, not only reading and writing files, but intelligently editing specific paragraphs or lines. Rather than generating entire files from scratch, Claude can apply targeted, multi-line edits (using its “Edit” and “MultiEdit” tools) to refine or improve existing content in a safe, sandboxed environment. I’ll walk through a production‑grade architecture where an agent crawls a website, captures content, and updates documentation files (Markdown, images, config) by refining sections, inserting new details, or restructuring existing docs. Along the way, I’ll explain how to configure tool‑level permissions (so Claude only edits allowed paths), how to use the SDK’s text-editor capabilities, and how to monitor and audit its file changes. By the end, users will understand how file‑system manipulation gives Claude a richer, more reliable “working memory”, letting it reason about and improve real documents over time, rather than just retrieving or generating content from a vector database.

- Event context: AI Tinkerers Montreal - January 2026 Meetup at Ateko — 2026-01-21 — Montreal
- Public talk page: https://montreal.aitinkerers.org/talks/rsvp_-GRVa1txV6o

### [Building The Agentic Layer](https://seattle.aitinkerers.org/talks/rsvp_rGiKGTRGC1M)

A high level walkthrough of how I use the Claude Code SDK to automate the entire SDLC for tasks in my code bases.

- Event context: AI Tinkerers Seattle Meetup: Dev Tools Track — November 3, 2025 — 2025-11-04 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_rGiKGTRGC1M

### [Multi-Agent Claude Code: Orchestrating Collaborative AI Coding Workflows](https://seattle.aitinkerers.org/talks/rsvp_ltB-QH4mt_I)

A desktop app and API for multi-agent collaborative development with Claude Code. Features include @agent-name mentions for direct task routing, automatic workflow orchestration across local and remote agents, and file-based coordination for complex development tasks. The system uses your existing Claude subscription (no API keys required) and supports dynamic agent configuration through a web UI. Demo will show the full workflow: from task decomposition by the orchestrator agent, to coordinated execution across specialized agents, to final code delivery with dependency management.

- Event context: AI Dev Tool Track — Seattle Meetup (Sept 8, 2025) — 2025-09-09 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_ltB-QH4mt_I

### [Multiple Claude Code collaborating as multi-agent](https://seattle.aitinkerers.org/talks/rsvp_fL0yMDL791U)

This talk showcases how we run multiple Claude Code agents using an open-source MCP (Model Context Protocol) setup — not as a demo, but as production infrastructure. Each agent runs inside a container, with file-mounted context, state, and config. They operate on real tasks (browser automation, LinkedIn browsing, competitive research) and coordinate via pure natural language prompts — just like humans chatting. All logs and exchanges are streamed into a shared chatroom, enabling review, human intervention, or agent-to-agent negotiation. No plugin wrappers. No forced APIs. Just prompt-level orchestration that works.

- Event context: Building AI Agents with Google Cloud AI — 2025-07-25 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_fL0yMDL791U

## Related Technologies

- [Claude](https://aitinkerers.org/technologies/claude) ([Markdown](https://aitinkerers.org/technologies/claude.md)) — 173 public demos
- [MCP](https://aitinkerers.org/technologies/mcp) ([Markdown](https://aitinkerers.org/technologies/mcp.md)) — 129 public demos
- [agent-browser](https://aitinkerers.org/technologies/agent-browser) ([Markdown](https://aitinkerers.org/technologies/agent-browser.md)) — 3 public demos
- [Agentic loop](https://aitinkerers.org/technologies/agentic-loop) ([Markdown](https://aitinkerers.org/technologies/agentic-loop.md)) — 1 public demo
- [Claude-3](https://aitinkerers.org/technologies/claude-3) ([Markdown](https://aitinkerers.org/technologies/claude-3.md)) — 110 public demos
- [Claude Code](https://aitinkerers.org/technologies/claude-code) ([Markdown](https://aitinkerers.org/technologies/claude-code.md)) — 214 public demos
- [Django](https://aitinkerers.org/technologies/django) ([Markdown](https://aitinkerers.org/technologies/django.md)) — 15 public demos
- [Edit](https://aitinkerers.org/technologies/edit) ([Markdown](https://aitinkerers.org/technologies/edit.md)) — 1 public demo
- [GitHub](https://aitinkerers.org/technologies/github) ([Markdown](https://aitinkerers.org/technologies/github.md)) — 74 public demos
- [GitHub API](https://aitinkerers.org/technologies/github-api) ([Markdown](https://aitinkerers.org/technologies/github-api.md)) — 6 public demos
- [GPT-4](https://aitinkerers.org/technologies/gpt-4) ([Markdown](https://aitinkerers.org/technologies/gpt-4.md)) — 529 public demos
- [Haiku](https://aitinkerers.org/technologies/haiku) ([Markdown](https://aitinkerers.org/technologies/haiku.md)) — 7 public demos
- [Llama-2](https://aitinkerers.org/technologies/llama-2) ([Markdown](https://aitinkerers.org/technologies/llama-2.md)) — 227 public demos
- [Mac Mini hardware](https://aitinkerers.org/technologies/mac-mini-hardware) ([Markdown](https://aitinkerers.org/technologies/mac-mini-hardware.md)) — 1 public demo
- [MultiEdit](https://aitinkerers.org/technologies/multiedit) ([Markdown](https://aitinkerers.org/technologies/multiedit.md)) — 1 public demo
- [Node](https://aitinkerers.org/technologies/node) ([Markdown](https://aitinkerers.org/technologies/node.md)) — 99 public demos
- [OpenAI API](https://aitinkerers.org/technologies/openai-api) ([Markdown](https://aitinkerers.org/technologies/openai-api.md)) — 520 public demos
- [Playwright](https://aitinkerers.org/technologies/playwright) ([Markdown](https://aitinkerers.org/technologies/playwright.md)) — 32 public demos