# JSON Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/json
> Markdown URL: https://aitinkerers.org/technologies/json.md
> Technology record last updated: 2026-02-25T06:29:39Z
> Generated: 2026-09-20T20:43:08Z

JSON (JavaScript Object Notation): A lightweight, language-independent data format for structured data interchange, built on universally supported key-value pairs and ordered arrays.

JSON is the standard for modern data exchange, leveraging human-readable text to transmit structured information. Its minimal syntax uses two core types: objects (unordered collections of name:value pairs, like {"user": "alpha"}) and arrays (ordered lists of values, like [1, 2, 3]). This structure, derived from JavaScript, ensures efficient parsing and generation across all major programming languages (e.g., Python, Java, Go). We deploy JSON extensively: it is the de facto payload for RESTful APIs and the preferred format for application configuration files (e.g., package.json). The current specification (RFC 8259) ensures consistent, high-scale interoperability.

- Official technology site: https://json.org
- Public AI Tinkerers demos and talks: 20
- Result page: 1 of 1

## Recent Public Talks and Demos

### [A context-aware interactive Piano tutor](https://atlanta.aitinkerers.org/talks/rsvp_Rf2oBe4bzK8)

VibeKeys is a piano learning app designed to test the boundaries of interactive AI education capabilities. I'l show an app demo where the student plays a section incorrectly, and how the voice AI agent is engaged to provide insight to the student through its multi-level agent orchestration and context engineering system. The app has an interactive context layer that allows low-latency responses for simple tool calling combined with more escalated handoff to a deep-reasoning "Music Education" agent. I could also show some of the cool AI tools and skills I use to build this project: - Project-Specific "Pattern Languages" (Christopher Alexander) to help guide decision-making - BAML (open-source AI wrapper infrastructure that turns all LLM calls into statically typed function calls). - My code review process which uses Reviewable.io to host the diffs and interactively communicate with the AI agent to address review issues.

- Event context: AI Tinkerers Atlanta: 2-Year Anniversary & Best Demos of the Year — 2026-07-30 — Atlanta
- Public talk page: https://atlanta.aitinkerers.org/talks/rsvp_Rf2oBe4bzK8

### [Robot Picks Grapes Based on Spatial World Model](https://seattle.aitinkerers.org/talks/rsvp_hwjgpFnwm2M)

We built a new model that is based on the human brain. Right now it can navigate and reason to play Pac Man but we are working on connecting it to a robot and having it pick grapes.

- Event context: AI Tinkerers Seattle Summer Bash — 2026-07-29 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_hwjgpFnwm2M

### [Turning a Hackathon Prototype into a Scalable Solution](https://saopaulo.aitinkerers.org/talks/rsvp_A_SzBK-d0uU)

JustaLuz is a mobile app that uses on-device capture plus Gemini vision to parse Brazilian utility bills, flag inconsistencies with deterministic rules and search-grounded tariff checks, and walk users through formal billing disputes. This addresses a recurring issue in Brazil: errors and confusing charges on electricity bills. The demo showed the end-to-end workflow in the app and illustrated how the team worked on evolving a hackathon prototype into a clear and scalable architecture.

- Event context: AI Tinkerers SP - Meetup de Maio - Kiro & AWS — 2026-05-28 — São Paulo
- Public talk page: https://saopaulo.aitinkerers.org/talks/rsvp_A_SzBK-d0uU

### [Domain-encoded system prompts: turning 15 years of accounting forensics into a live Claude diagnostic tool](https://montreal.aitinkerers.org/talks/rsvp_x6xfCYXZsaM)

I built a structured diagnostic engine that ingests raw ADP Workforce Now payroll exports and NetSuite GL data, routes them through a domain-encoded Claude system prompt, and returns a JSON-structured root cause analysis — including the specific fix — for recurring payroll-GL discrepancies that human reviewers consistently miss. The demo runs live against the Anthropic Messages API. The input is a real payroll pattern: employer burden accounts (CPP, EI, CNESST, RQAP) posting at exactly 0.500x expected in every quarter-end month for three years. The system prompt carries the diagnostic heuristics — check ratio before amount, look for N-period calendar patterns, separate gross wages from employer burden — that normally live inside a practitioner's head. Claude doesn't discover these rules. They're injected. The output is a structured JSON object with root_cause, confidence_pct, adp_fix_steps[], correcting_je_description, and cfo_memo.

- Event context: AI Tinkerers Montreal - March Demo Night — 2026-03-24 — Montreal
- Public talk page: https://montreal.aitinkerers.org/talks/rsvp_x6xfCYXZsaM

### [Prompt Stretching: Applying Linear Transformations and Recursive Decomposition to Prompts](https://la.aitinkerers.org/talks/rsvp_nkNEynHMFvs)

I'll walk through an open source Python library I'm working on where LLMs generate other LLMs as reusable JSON configs. The core demo is "prompt stretching" — a pipeline that takes a single user prompt and fans it out through a family tree of transformations (inverse, abstract up/down, reflection, rotation, shear, scaling, recursion, dimension shift). Then I'll also demo subprompting — recursively decomposing a prompt into a tree of subprompts at configurable depth and branching factor (e.g. n=[4,3,2]), with concurrent execution, so anyone can build their own deep research agent. Finally, all the transformed/decomposed prompt variants and their responses can be pairwise evaluated.

- Event context: AI Tinkerers LA: March Builder Meetup & Live Demos — 2026-03-20 — Los Angeles
- Public talk page: https://la.aitinkerers.org/talks/rsvp_nkNEynHMFvs

### [Coding with AI: What Works & What Doesn't](https://manchester-nh.aitinkerers.org/talks/rsvp_WlqmhQoEJnw)

We all know AI dev tools are powerful, but they can also be incredibly frustrating if you use them wrong. In this session, I'll share my personal "dos and don'ts" for navigating the current landscape of AI assistants. We'll cover the right (and wrong) ways to use different tools, and how to avoid the common traps that actually slow you down. Finally, I’ll show you what happens when you build your own rules: a live demo of a self-learning agent I built with Claude Code and mcp.json that records sessions, writes its own memories, and gets smarter as I code.

- Event context: AI Tinkerers Manchester (Bedford), NH - March 2026 Meetup — 2026-03-18 — Manchester NH
- Public talk page: https://manchester-nh.aitinkerers.org/talks/rsvp_WlqmhQoEJnw

### [Poor Man’s Long Running Multi Hour Coding Agent](https://seattle.aitinkerers.org/talks/rsvp_2_7aWPHU1uc)

In December 2025, I was running out of time. In an act of desperation and after reading Anthropix's blog post or paper about long-running agentic harnesses, I attempted to replicate what they described in their paper and was able to achieve over a long weekend the completion of a feature done enough and well enough that I could ship and meet my deadline. The outcome of this was over that weekend I completely maxed out my Claude Max subscription and then on top of that I incurred an overage of $700.

- Event context: AI Tinkerers Seattle Meetup: Dev Tools Track — February 9th, 2026 — 2026-02-10 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_2_7aWPHU1uc

### [Syntropix: The AI Workforce Powering the Next Generation of SMEs](https://london.aitinkerers.org/talks/rsvp_bFCMhts7fzk)

Syntropix is building a proprietary AI-native database that consolidates fragmented business data from ERPs, emails, spreadsheets, and external systems into a structured, queryable source of truth. This database provides contextual grounding for our agentic automation framework: role-aware AI agents can retrieve, interpret, and act on data with fine-grained permissions, precision, and traceability. By coupling a context-rich datastore with intelligent agents, we enable SMEs to automate complex workflows—bank reconciliation, procurement, customer service—with reliability and real-time adaptability that traditional RPA, SaaS, or off-the-shelf AI cannot achieve.

- Event context: AI Tinkerers x MongoDB - Demopalooza — 2025-10-07 — London
- Public talk page: https://london.aitinkerers.org/talks/rsvp_bFCMhts7fzk

### [Artixa](https://seattle.aitinkerers.org/talks/rsvp_3aSxCgRSYWc)

Artixa is an AI-driven artwork recognition and education app. Using Artixa on a mobile phone, gallery owners, gallery patrons, or museum visitors can take a picture of an artwork and be presented with detailed information about the work and artist. Should they wish to learn more, users can use the chat interface to ask Artixa questions about the work, artist, movement, and other related topics.

- Event context: AI Tinkerers Seattle - February 2025 Meetup — 2025-02-22 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_3aSxCgRSYWc

### [Self Correcting Data Extractor](https://mumbai.aitinkerers.org/talks/rsvp_Zr3pPQHK26M)

Coding agents collaboratively write and debug Python code to convert bank statements from any bank into standardized JSON. Using various PDF-to-text libraries, agents iteratively refine their code to handle errors. Successful code is stored for future processing of statements from the same bank.

- Event context: 🚀 AI Tinkerers Mumbai Chapter: Community Social Event! — 2025-01-25 — Mumbai
- Public talk page: https://mumbai.aitinkerers.org/talks/rsvp_Zr3pPQHK26M

### [Cooktok](https://montreal.aitinkerers.org/talks/rsvp_5t8RKHulk5Q)

I made a really basic AI app to help me turn TikTok videos into recipes so I can cook with my husband. It takes the captions and description and turns it into a JSON object so I can print it out and cook it.

- Event context: AI Tinkerers - Montreal Inaugural Meetup (October) — 2024-10-29 — Montreal
- Public talk page: https://montreal.aitinkerers.org/talks/rsvp_5t8RKHulk5Q

### [Tables and AI](https://seattle.aitinkerers.org/talks/rsvp_ZBSLAytd6ms)

In his classic book Code Complete, Steve McConnell dedicates a whole chapter to Table Driven Methods. A Table Driven method is the use of lookups in lieu of complicated algorithms. It offers simplicity as an alternative to complexity; immediate modifiability over hard coding. With AI, this paradigm is more relevant than ever before and in fact essential when prioritizing the role of people in defining the rules an LLM should follow. With tables, literally worksheets, stakeholders can manage, discuss, and evaluate rules for how an AI responds and the structure of the response(think JSON data structures). For a solo developer, tables offer a cell for every thought and an opportunity for every thought to be thoroughly thought through. Technically, tables allow us to categorize, weigh, engineer prompts, and then manage the execution of individual and grouped prompts. Because LLMs are imperfect in what they return, we can use a table to compare what was produced with what was expected. A lot of value in a near forgotten paradigm, whose time has come to shine.

- Event context: AI Tinkerers - Seattle - October 2024 Meetup — 2024-10-25 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_ZBSLAytd6ms

### [AI WebScraper](https://abu-dhabi.aitinkerers.org/talks/rsvp_eK6IaDeOIVI)

An AI webscraping application using GenAI where a user can simply put the URL of a website that he wants to scrape and then input keywords to scrape. For example, if the website is about properties then the keywords would be "image, location, price, area" then you have 3 options where you can download it as JSON, CSV, or Markdown

- Event context: AI Tinkerers - Abu Dhabi Inaugural Meetup (September 2024) — 2024-09-10 — Abu Dhabi
- Public talk page: https://abu-dhabi.aitinkerers.org/talks/rsvp_eK6IaDeOIVI

### [Cut the red tape when asking for a work permit from Dubai Asset Management](https://abu-dhabi.aitinkerers.org/talks/rsvp_gmVoO2LHnxU)

If you are _lucky_ enough to live in a community managed by Dubai Properties you know that even to have furniture delivered you need to fill in a pdf form and send it by email. I plan to show a way to automate this process by explaining what you want to do in natural language and have the form filled in for you. If I manage I will also show how to integrate it with an email client to create the email with the filled in PDF form as an attachment At the moment I am using llama3.1 but I might try with other models as well. I plan to explain the tools I used and the issues I faced, especially with having the LLM outputting structured (json) data

- Event context: AI Tinkerers - Abu Dhabi Inaugural Meetup (September 2024) — 2024-09-10 — Abu Dhabi
- Public talk page: https://abu-dhabi.aitinkerers.org/talks/rsvp_gmVoO2LHnxU

### [AI-Driven Triage Automation: Voice-to-Text with Groq for Emergency Care](https://dubai.aitinkerers.org/talks/rsvp_RQGteQRFzsM)

This session will dive into how Groq's AI models can be applied to automate triage documentation in emergency care. Using a real-world implementation, we’ll demonstrate how paramedic or doctor statements can be transcribed using the Whisper-large-v3 model, and automatically converted into structured JSON data for patient records. The code leverages Groq’s transcription capabilities and LLaMA3-70B for generating structured data that populates triage forms, saving critical time in emergencies. The presentation includes a detailed walkthrough of the code, covering everything from uploading the audio file to generating the structured output, along with a live demo.

- Event context: AI Tinkerers - Dubai Inaugural Meetup (September) — 2024-09-07 — Dubai
- Public talk page: https://dubai.aitinkerers.org/talks/rsvp_RQGteQRFzsM

### [LLM powered email phishing simulation](https://dublin.aitinkerers.org/talks/rsvp_2NdkJZ-bnY4)

Companies use phishing simulation to train employees and prevent the loss of company assets through this malicious but very effective technique, llm powered lure generation show cases just how quick easy and personalised phishing has become and how it can be produced at scale

- Event context: AI Tinkerers - Dublin Inaugural Meetup (September) — 2024-09-05 — Dublin
- Public talk page: https://dublin.aitinkerers.org/talks/rsvp_2NdkJZ-bnY4

### [Everchanging quest, a Roguelike in which the LLM controls the game](https://paris.aitinkerers.org/talks/rsvp_EuiJ473-gP0)

It is a roguelike in which the layout is passed to the LLM as Markdown. The LLM is then responsible for generating a JSON that contains an objective to fulfill as well as the necessary objects to do so.

- Event context: AI Tinkerers - Paris Inaugural Meetup featuring Mistral AI — 2024-09-03 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_EuiJ473-gP0

### [NuExtract - a foundation model for Structured Extraction](https://paris.aitinkerers.org/talks/rsvp_-BOqa4DWNMs)

In this talk, I will present the NuExtract research project. NuExtract is a lightweight open-source foundation model specialized in the task of structured extraction - allowing to extract arbitrarily complex information from text and turn it into structured data in the form of a JSON file. In a nutshell, NuExtract is trained on an LLM-annotated dataset carefully crafted to mitigate hallucinations. I will explain and demonstrate how we created the first version of NuExtract, focusing on the technical choices we made and the various solutions we found to make this work. I will also discuss about possible future improvements. More about the project: https://numind.ai/blog/nuextract-a-foundation-model-for-structured-extraction

- Event context: AI Tinkerers - Paris Inaugural Meetup featuring Mistral AI — 2024-09-03 — Paris
- Public talk page: https://paris.aitinkerers.org/talks/rsvp_-BOqa4DWNMs

### [SumSpot App](https://seattle.aitinkerers.org/talks/rsvp_WbRcHFSoum8)

A Safari extension on iPhone that summarizes [news] articles. (40 seconds demo video)

- Event context: AI Tinkerers Seattle - May 2024 — 2024-05-21 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_WbRcHFSoum8

### [Enhancing Application Data and Logic with OpenAI Functions](https://chicago.aitinkerers.org/talks/rsvp_uu5a4Xh3nks)

OpenAI Functions provides a means of translating generative AI responses into application processable JSON, bridging the gap of generative AI from text to data. Let’s see some ways OpenAI Functions can be leveraged to enhance a simple database web application!

- Event context: Chicago AI Tinkerers 2024 Kickoff — 2024-01-23 — Chicago
- Public talk page: https://chicago.aitinkerers.org/talks/rsvp_uu5a4Xh3nks

## Related Technologies

- [Claude Code](https://aitinkerers.org/technologies/claude-code) ([Markdown](https://aitinkerers.org/technologies/claude-code.md)) — 214 public demos
- [LLM](https://aitinkerers.org/technologies/llm) ([Markdown](https://aitinkerers.org/technologies/llm.md)) — 123 public demos
- [OpenAI API](https://aitinkerers.org/technologies/openai-api) ([Markdown](https://aitinkerers.org/technologies/openai-api.md)) — 520 public demos
- [AI](https://aitinkerers.org/technologies/ai) ([Markdown](https://aitinkerers.org/technologies/ai.md)) — 55 public demos
- [Gemini](https://aitinkerers.org/technologies/gemini) ([Markdown](https://aitinkerers.org/technologies/gemini.md)) — 188 public demos
- [Markdown](https://aitinkerers.org/technologies/markdown) ([Markdown](https://aitinkerers.org/technologies/markdown.md)) — 24 public demos
- [PDF](https://aitinkerers.org/technologies/pdf) ([Markdown](https://aitinkerers.org/technologies/pdf.md)) — 10 public demos
- [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) — 662 public demos
- [ADP Workforce Now](https://aitinkerers.org/technologies/adp-workforce-now) ([Markdown](https://aitinkerers.org/technologies/adp-workforce-now.md)) — 1 public demo
- [AI agents](https://aitinkerers.org/technologies/ai-agents) ([Markdown](https://aitinkerers.org/technologies/ai-agents.md)) — 35 public demos
- [AI-native databases](https://aitinkerers.org/technologies/ai-native-databases) ([Markdown](https://aitinkerers.org/technologies/ai-native-databases.md)) — 1 public demo
- [Anthropic API](https://aitinkerers.org/technologies/anthropic-api) ([Markdown](https://aitinkerers.org/technologies/anthropic-api.md)) — 66 public demos
- [Anthropic Messages API](https://aitinkerers.org/technologies/anthropic-messages-api) ([Markdown](https://aitinkerers.org/technologies/anthropic-messages-api.md)) — 2 public demos
- [Antigravity](https://aitinkerers.org/technologies/antigravity) ([Markdown](https://aitinkerers.org/technologies/antigravity.md)) — 6 public demos
- [Apache PDFBox](https://aitinkerers.org/technologies/apache-pdfbox) ([Markdown](https://aitinkerers.org/technologies/apache-pdfbox.md)) — 1 public demo
- [Apache Tika](https://aitinkerers.org/technologies/apache-tika) ([Markdown](https://aitinkerers.org/technologies/apache-tika.md)) — 1 public demo
- [API](https://aitinkerers.org/technologies/api) ([Markdown](https://aitinkerers.org/technologies/api.md)) — 19 public demos
- [BAML](https://aitinkerers.org/technologies/baml) ([Markdown](https://aitinkerers.org/technologies/baml.md)) — 12 public demos