# MCP Projects at AI Tinkerers
> Canonical HTML: https://aitinkerers.org/technologies/mcp?page=2
> Markdown URL: https://aitinkerers.org/technologies/mcp.md?page=2
> Technology record last updated: 2026-09-18T15:14:00Z
> Generated: 2026-09-22T18:42:54Z
MCP is the open-source standard for securely connecting AI agents (like LLMs) to external tools, data, and enterprise workflows.
The Model Context Protocol (MCP) functions as a standardized integration layer: think of it as a USB-C port for AI applications. Developed and open-sourced by Anthropic, this protocol allows large language models (LLMs) to access real-time context and execute actions via external tools like GitHub, Jira, or proprietary databases . It uses a simple JSON-RPC interface to define tools, schemas, and endpoints, which enables AI agents to perform complex, state-changing tasks—such as creating a GitHub issue or running a test script—rather than just generating text . MCP is essential for building agentic AI systems that can autonomously pursue goals and operate within defined safety and permission boundaries .
- Official technology site: https://modelcontextprotocol.io/
- Public AI Tinkerers demos and talks: 129
- Result page: 2 of 6
## Recent Public Talks and Demos
### [ArgosBrain pushed Opus 4.7 from 87.6% to 95% on SWE-bench Verified](https://paris.aitinkerers.org/talks/rsvp_xM-68JzVhpc)
ArgosBrain — structural engine for AI coding agents. Opus 4.7 leads the GA SWE-bench Verified leaderboard at 87.6%. We're aiming for ~95% — by amplifying (not replacing) grep and Read with ArgosBrain. Argos delivers structural facts — every symbol, every caller, every type — in under 50ms at $0 per query. Grep and Read get faster and sharper because they know exactly where to search. We're publishing the full head-to-head: Opus 4.7 vanilla vs Opus 4.7 + ArgosBrain on SWE-bench Verified. AI agents like Claude Code and Cursor re-read your codebase every session, burning 73% of their tokens on redundant file reads. ArgosBrain is a local Rust engine that indexes your code into a unified graph (SCIP + LSP + tree-sitter) and serves it to agents via MCP — sub-50ms symbol lookups, $0 per query, zero hallucinated symbols. Live tested on Kubernetes (~2M LOC), VS Code, and OpenClaw. Per-task cost drops from $0.40 to $0.04 in our benchmarks.
- Event context: AI Tinkerers Paris feat Scalingo: Conversational DevOps & AI Infrastructure — 2026-05-21 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_xM-68JzVhpc
### [Turning Claude into a Transactional Lawyer: Building a Legal Contract RAG System as an MCP Server](https://raleigh.aitinkerers.org/talks/rsvp_s4Fh6KIHlHg)
I built a RAG agent for answering questions about contracts, then converted it to an MCP server that gives Claude the ability to search, analyze, and compare a corpus of 500+ legal contracts using retrieval-augmented generation. In the demo, I'll connect it to Claude Desktop and show multi-turn contract Q&A — asking about termination clauses, comparing provisions across agreements, and testing whether the system resists hallucinating clauses that don't exist.
- Event context: AI Tinkerers Raleigh Meetup — May 6, 2026 — 2026-05-06 — Raleigh
- Public talk page: https://raleigh.aitinkerers.org/talks/rsvp_s4Fh6KIHlHg
### [Methodologies for the Optimization of Developers' Workflow with AI Agents](https://pereira.aitinkerers.org/talks/rsvp_9_pwCnocOjU)
This repository is an architectural and educational guide designed to enhance Artificial Intelligence agents through the implementation of Plugins, MCP servers, and Custom Skills, built upon an open‑source software that enables fast connections between MCP servers and OpenCode. The workflow for software development has changed significantly with the arrival of artificial intelligence. However, the advances have been very rapid and numerous, making it difficult to keep up. For quite some time, I have observed that new tools and approaches constantly emerge, and most people —myself included— struggle to keep pace. Many resort to vibe coding, obtaining only average results. In response to this phenomenon, I identified that there are different ways to improve the daily use of agents. That is why I created this repository, conceived from the perspective of a student for other students. The goal is to leverage the wide variety of agents with powerful free tiers and explore how to enhance them, whether free or paid, through different methodologies. The ultimate aim is to move beyond vibe coding and instead embrace a more solid practice: becoming an AI Engineer.
- Event context: AI Tinkerers Pereira: Innovando con IA desde el SENA — 2026-04-30 — Pereira
- Public talk page: https://pereira.aitinkerers.org/talks/rsvp_9_pwCnocOjU
### [Watch a Sales Guy Ship Agentic Infrastructure](https://toronto.aitinkerers.org/talks/rsvp_UMP1pQfK9Bo)
Human commerce runs on trust. Agentic commerce will run on confidence. Disclose Framework is an open standard for publishing the operational signals autonomous systems need to evaluate businesses with confidence. This demo will focus on commerce. It shows an MCP server that reads a merchant's /.well-known/disclose file and returns structured operational data, including return rate, fulfillment accuracy, and chargeback ratio, so an agent can evaluate a seller on verifiable facts rather than reviews or price alone.
- Event context: AI Tinkerers Toronto - April 2026 - hosted by Shopify — 2026-04-29 — Toronto
- Public talk page: https://toronto.aitinkerers.org/talks/rsvp_UMP1pQfK9Bo
### [Learnings of integrating OpenClaw & Local LLM into my Home Assistant instance](https://hong-kong.aitinkerers.org/talks/rsvp_4b4SPWWvW4A)
Does OpenClaw elevate my smart home experience on top of the already powerful Home Assistant? By leveraging the OpenClaw add-on in Home Assistant and connect it to its MCP Server, I managed to control part of my smart home via Telegram. I explored the capabilities of the latest local LLMs (or "SLMs"?), such as Gemma 4 E4B, Qwen 3.5 9B, GPT-OSS-20B etc., and experienced some hit-and-misses along the way. I will share what worked well, what worked, and what didn't.
- Event context: AI Tinkerers Hong Kong at AWS: Agentic AI in Action (April) — 2026-04-29 — Hong Kong
- Public talk page: https://hong-kong.aitinkerers.org/talks/rsvp_4b4SPWWvW4A
### [From Messy Inputs to Reliable Community Data: Human-in-the-Loop AI](https://denver-boulder.aitinkerers.org/talks/rsvp_EuM3A-aBwJ4)
I built Colorado Songwriters Collective, a human-in-the-loop AI workflow that converts messy event/member inputs (DMs, forms, and web pages) into structured, publish-ready community data. In the demo, I’ll show the ingestion pipeline, review checkpoints, and how we keep updates accurate before they go live.
- Event context: AI Tinkerers Denver - April Meeting — 2026-04-27 — Denver
- Public talk page: https://denver-boulder.aitinkerers.org/talks/rsvp_EuM3A-aBwJ4
### [Hive: Local-First AI Gateway](https://bogota.aitinkerers.org/talks/rsvp_ukV86hfl3Q0)
Explore building a local-first AI agent gateway with Bun and SQLite. Learn to orchestrate autonomous agents across multiple communication channels, manage tools, and switch LLM providers seamlessly.
- Event context: AI Tinkerers Bogotá: Demos for builders — 2026-04-23 — Bogotá
- Public talk page: https://bogota.aitinkerers.org/talks/rsvp_ukV86hfl3Q0
### [Deterministic tools, Non-Deterministic orchestration — Building a production AI Agent for auto parts lookup](https://poland.aitinkerers.org/talks/rsvp_8D7UFMN443Y)
An AI agent that takes a car's registration number and returns purchasable replacement parts, solving a problem of a fragmented Polish aftermarket where no single API covers all brands. The system splits the problem into deterministic source selection (which API to call, in what order, per brand) and an LLM-orchestrated agent asks clarifying questions when data is ambiguous, and combines partial results from multiple sources.
- Event context: AI Tinkerers Poland - Meetup in Gdańsk #1 — 2026-04-23 — Poland
- Public talk page: https://poland.aitinkerers.org/talks/rsvp_8D7UFMN443Y
### [Docker Sandboxes](https://paris.aitinkerers.org/talks/rsvp_q-ziE0PuCC8)
AI is everywhere and easier to use. We think it can be even simpler. With AI, we either write a prompt, use a pre-built agent, or code complex agents. We think there's room for ad hoc agents written in seconds, without code. Why? Because AI is often useful for tasks never seen before, and it's a superpower to assemble models, prompts, and tools into a team of agents created for that task.
- Event context: AI Tinkerers Paris: Docker Agentic Workflows (Devoxx Kickoff) — 2026-04-21 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_q-ziE0PuCC8
### [Docker Agent: Simplifying AI Agent Creation](https://paris.aitinkerers.org/talks/rsvp_5hoKr2cgNMw)
AI is everywhere and easier to use. We think it can be even simpler. With AI, we either write a prompt, use a pre-built agent, or code complex agents. We think there's room for ad hoc agents written in seconds, without code. Why? Because AI is often useful for tasks never seen before, and it's a superpower to assemble models, prompts, and tools into a team of agents created for that task.
- Event context: AI Tinkerers Paris: Docker Agentic Workflows (Devoxx Kickoff) — 2026-04-21 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_5hoKr2cgNMw
### [Docker Agent: Simplifying AI Agent Creation](https://paris.aitinkerers.org/talks/rsvp_QMDTiAwXhfE)
AI is everywhere and easier to use. We think it can be even simpler. With AI, we either write a prompt, use a pre-built agent, or code complex agents. We think there's room for ad hoc agents written in seconds, without code. Why? Because AI is often useful for tasks never seen before, and it's a superpower to assemble models, prompts, and tools into a team of agents created for that task.
- Event context: AI Tinkerers Paris: Docker Agentic Workflows (Devoxx Kickoff) — 2026-04-21 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_QMDTiAwXhfE
### [Harness: A Multi-Agent Claude Orchestrator That Builds and QA's Games Autonomously](https://singapore.aitinkerers.org/talks/rsvp_RwBQJVJwoLk)
Harness is a game-agnostic multi-agent orchestrator that uses the Claude Agent SDK to autonomously build and test games end-to-end. It reads a feature-list.json from any target game project, spawns a role-specific generator agent (gameplay / UI-art / audio) to implement each feature with full tool access (Read, Write, Edit, Bash, Glob, Grep + Godot MCP), then spawns a separate evaluator agent that launches the game, classifies each test step as DATA / VISUAL / BOTH, captures screenshots, and emits pass/fail verdicts. A watchdog loop detects stuck sessions and a pre-flight smoke test validates tool permissions before real work begins. Everything streams live to a single-process Fastify + React dashboard over WebSockets — you watch agents ship and break features in real time. For the demo I'll drive Harness against a live Godot project, show the full loop (pick → implement → evaluate → retry), and walk through how session records, screenshots, and eval reports are persisted back to the target repo.
- Event context: AI Tinkerers Singapore: The Agentic Future & Dev/Eng Workflows — 2026-04-21 — Singapore
- Public talk page: https://singapore.aitinkerers.org/talks/rsvp_RwBQJVJwoLk
### [CodeDB: Building a Code Intelligence Server That Cuts Agent Token Usage](https://singapore.aitinkerers.org/talks/rsvp_YyFDIYV6hPs)
CodeDB is a code intelligence server for AI agents that indexes a codebase once and serves structured answers through MCP, HTTP, and CLI instead of forcing models to repeatedly scan raw files. In this demo, I show how CodeDB uses structural indexing, trigram search, inverted word lookups, dependency graphs, and snapshots to help agents find only the most relevant code context with much lower latency and dramatically fewer tokens.
- Event context: AI Tinkerers Singapore: The Agentic Future & Dev/Eng Workflows — 2026-04-21 — Singapore
- Public talk page: https://singapore.aitinkerers.org/talks/rsvp_YyFDIYV6hPs
### [Adaptive Tool Orchestration with On-Demand Discovery and Code Mode Execution](https://ho-chi-minh-city.aitinkerers.org/talks/rsvp_ha31N9dlK_A)
A unified MCP transform that replaces static tool catalogs with intelligent, on-demand discovery while enabling dynamic orchestration through code mode execution. Instead of exposing every tool upfront, the system allows the model to search and resolve tools contextually, then generate and execute controlled C# workflows to chain calls efficiently. Implemented using the GitHub Copilot C#/.NET SDK and the MCP .NET SDK, this approach minimizes prompt overhead, reduces round-trips, and enables scalable, runtime tool composition within enterprise-grade .NET environments.
- Event context: AI Tinkerers Ho Chi Minh City: AI Coding Agents & Orchestrators — 2026-04-18 — Ho Chi Minh City
- Public talk page: https://ho-chi-minh-city.aitinkerers.org/talks/rsvp_ha31N9dlK_A
### [ClawForce: Let Claude Spawn People to Do Things at Scale](https://austin.aitinkerers.org/talks/rsvp_24BdA7wnyQI)
I built a lite MCP that let's you expose commissioning real people to do anything in the physical world from an AI agentic workflow. Think Terraform, but instead of provisioning servers, you're provisioning people with Door Dash, Uber Direct, TaskRabbit, etc.
- Event context: AI Tinkerers Austin: April 2026 Demo Night — 2026-04-16 — Austin
- Public talk page: https://austin.aitinkerers.org/talks/rsvp_24BdA7wnyQI
### [Bugs Fix Themselves Now](https://upstate-ny.aitinkerers.org/talks/rsvp_75Gx_cCu2T8)
I built an agent loop that connects an error tracker to a coding agent so bugs go from "detected" to "fixed" without me context switching. Live demo: I'll trigger errors in a real app, then let Claude Code query them via MCP and generate the fix - all in the terminal.
- Event context: AI Tinkerers Upstate NY: Agents and Automation — 2026-04-14 — Upstate NY
- Public talk page: https://upstate-ny.aitinkerers.org/talks/rsvp_75Gx_cCu2T8
### [Independent Agents to Hive Mind - agent/chatbot/human collaboration via hiveWiki.ai](https://upstate-ny.aitinkerers.org/talks/rsvp_l5dVgas09AQ)
Wiki-based agent collaboration tool accessed by chatbots and agents via MCP - and by humans via the wiki artifact. Other systems try to workflow agents like employees. I'll demonstrate Claude Chat, a human, and Cursor Agent collaborating as experts using a shared, human-usable context to design and deploy a feature.
- Event context: AI Tinkerers Upstate NY: Agents and Automation — 2026-04-14 — Upstate NY
- Public talk page: https://upstate-ny.aitinkerers.org/talks/rsvp_l5dVgas09AQ
### [Build MCP apps the easy way](https://zurich.aitinkerers.org/talks/rsvp_yY7TrFFu3wo)
mcp-use is one of the most popular frameworks for building MCP servers and MCP apps. At Manufact we love open source and we are creating the open source developer platform for MCP. in this demo, I will show how you can use the CLI or Manufact Vibe to create an MCP app in a couple of minutes, either through code or with Vibe coding, and then show how you can submit your app to the OpenAI marketplace
- Event context: AI Tinkerers Zurich April 9th — 2026-04-09 — Zürich
- Public talk page: https://zurich.aitinkerers.org/talks/rsvp_yY7TrFFu3wo
### [The Zero-Partners VC - AI Native VC running on AI Agents](https://zurich.aitinkerers.org/talks/rsvp_cPNM9rUu4j8)
We are www.ellipsis-venture.com - 2 GPs who are AI Builders (x-Google, x-founders, x-Apple) who run a fund without employees. We built an agentic system that runs everything - sourcing, due diligence, score cards and memos, marketing, Investors Relations, Ops, etc.
- Event context: AI Tinkerers Zurich April 9th — 2026-04-09 — Zürich
- Public talk page: https://zurich.aitinkerers.org/talks/rsvp_cPNM9rUu4j8
### [Your AI is Two Versions Behind](https://zurich.aitinkerers.org/talks/rsvp_k0YRpX9Yc4E)
A simple, repeatable system for filling in what AI models don’t know. Every model has a training cutoff, and everything after that date is a blind spot. I built structured markdown folders that capture what changed in a language or framework since the cutoff, and feed them directly into AI coding tools as project context. My first implementation covers Go 1.25 and 1.26, but the approach works for anything. I’ll demo the folder structure, show the before/after difference in AI output, and walk through gobot, a zero-dependency Go app built entirely with AI that had the missing knowledge loaded.
- Event context: AI Tinkerers Zurich April 9th — 2026-04-09 — Zürich
- Public talk page: https://zurich.aitinkerers.org/talks/rsvp_k0YRpX9Yc4E
### [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.   
- Event context: AI Tinkerers Denver Meetup - March 25, 2026 — 2026-03-25 — Denver
- Public talk page: https://denver-boulder.aitinkerers.org/talks/rsvp_ZkszEZey4aY
### [Turning a Textbook into an Architecture Consultant with MCP and Knowledge Graphs](https://denver-boulder.aitinkerers.org/talks/rsvp_pLy67GeNPpk)
An MCP server that extracts a knowledge graph from *Agentic Architectural Patterns for Building Multi-Agent Systems* (Arsanjani & Bustos, Packt 2026 [] ) — 138 concepts, 462 relationships, 786 sections — and uses it to run structured architecture consultations inside Claude Code. You point it at a codebase, it matches concepts via embeddings, spawns parallel subagents to traverse the graph for prerequisites/conflicts/alternatives, scores maturity against the book's Chapter 12 rubric (7 categories,36 patterns), and renders an interactive HTML report with before/after architecture diagrams. I'll walk through a pre-run consultation against OpenAI's Financial Research Agent (https://github.com/openai/openai-agents-python/tree/main/examples/financial_research_agent), scrolling through the real terminal output to show each phase: concept matching, scatter-gather subagent traversal, pattern assessment logging, deterministic scoring, and server-side report rendering.
- Event context: AI Tinkerers Denver Meetup - March 25, 2026 — 2026-03-25 — Denver
- Public talk page: https://denver-boulder.aitinkerers.org/talks/rsvp_pLy67GeNPpk
### [Agentes Interesantes a Infraestructura Confiable: Observabilidad, Gobernanza y Control para Sistemas Multi-Agente](https://pereira.aitinkerers.org/talks/rsvp_ll14d7Y1qrI)
Un framework de cuatro capas (observabilidad, gobernanza, seguridad operativa y regulación) para integrar sistemas agénticos responsablemente dentro de arquitecturas empresariales. Construido desde la experiencia de implementar IA en empresas en Chile, incluyendo el diseño de endpoints de reglas, cadenas de responsabilidad para agentes autónomos y sistemas de auditoría continua. Actualmente en producción para el validado con la gobernanza legal diseñada junto a asesores jurídicos especializados en IA.
- Event context: AI Tinkerers Pereira — The Builders Lab — 2026-03-25 — Pereira
- Public talk page: https://pereira.aitinkerers.org/talks/rsvp_ll14d7Y1qrI
### [Eliminating Waste in the AI Augmented SDLC](https://nashville.aitinkerers.org/talks/rsvp_uMKhslYmRfU)
When AI can use the same tools you use - Jira, GitHub, Sentry - the bottleneck in software development shifts from writing code to everything around the code. I applied lean manufacturing principles (Toyota Production System) to my own AI-augmented workflow and systematically removed entire categories of waste: manual ticket management, error triage, deployment monitoring, and ambiguous specifications that lead to expensive rework. I’ll walk through the specific patterns - MCP tool handoff, Gherkin specs as agent contracts, pre-flight checklists derived from PR review audits - and show how each elimination is permanent and reveals the next layer.
- Event context: AI Tinkerers Nashville – Software Development in the Age of AI — 2026-03-25 — Nashville
- Public talk page: https://nashville.aitinkerers.org/talks/rsvp_uMKhslYmRfU
## Related Technologies
- [Claude Code](https://aitinkerers.org/technologies/claude-code) ([Markdown](https://aitinkerers.org/technologies/claude-code.md)) — 216 public demos
- [TypeScript](https://aitinkerers.org/technologies/typescript) ([Markdown](https://aitinkerers.org/technologies/typescript.md)) — 205 public demos
- [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) — 662 public demos
- [Claude](https://aitinkerers.org/technologies/claude) ([Markdown](https://aitinkerers.org/technologies/claude.md)) — 174 public demos
- [OpenAI API](https://aitinkerers.org/technologies/openai-api) ([Markdown](https://aitinkerers.org/technologies/openai-api.md)) — 520 public demos
- [Gemini](https://aitinkerers.org/technologies/gemini) ([Markdown](https://aitinkerers.org/technologies/gemini.md)) — 188 public demos
- [Docker](https://aitinkerers.org/technologies/docker) ([Markdown](https://aitinkerers.org/technologies/docker.md)) — 147 public demos
- [LangChain](https://aitinkerers.org/technologies/langchain) ([Markdown](https://aitinkerers.org/technologies/langchain.md)) — 445 public demos
- [Node](https://aitinkerers.org/technologies/node) ([Markdown](https://aitinkerers.org/technologies/node.md)) — 99 public demos
- [Playwright](https://aitinkerers.org/technologies/playwright) ([Markdown](https://aitinkerers.org/technologies/playwright.md)) — 32 public demos
- [GitHub](https://aitinkerers.org/technologies/github) ([Markdown](https://aitinkerers.org/technologies/github.md)) — 74 public demos
- [LLM](https://aitinkerers.org/technologies/llm) ([Markdown](https://aitinkerers.org/technologies/llm.md)) — 123 public demos
- [Cursor](https://aitinkerers.org/technologies/cursor) ([Markdown](https://aitinkerers.org/technologies/cursor.md)) — 65 public demos
- [Next](https://aitinkerers.org/technologies/next) ([Markdown](https://aitinkerers.org/technologies/next.md)) — 186 public demos
- [GitHub Actions](https://aitinkerers.org/technologies/github-actions) ([Markdown](https://aitinkerers.org/technologies/github-actions.md)) — 20 public demos
- [HTTP](https://aitinkerers.org/technologies/http) ([Markdown](https://aitinkerers.org/technologies/http.md)) — 14 public demos
- [OpenAI](https://aitinkerers.org/technologies/openai) ([Markdown](https://aitinkerers.org/technologies/openai.md)) — 112 public demos
- [PostgreSQL](https://aitinkerers.org/technologies/postgresql) ([Markdown](https://aitinkerers.org/technologies/postgresql.md)) — 106 public demos
## More Results
- Previous: https://aitinkerers.org/technologies/mcp.md
- Next: https://aitinkerers.org/technologies/mcp.md?page=3