# Gradio Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/gradio
> Markdown URL: https://aitinkerers.org/technologies/gradio.md
> Technology record last updated: 2026-09-18T14:19:40Z
> Generated: 2026-09-23T03:50:08Z

Gradio is the open-source Python library for rapidly building and sharing interactive web UIs for any machine learning model or Python function.

Gradio is the essential tool for data scientists and ML engineers: it turns any Python function (including TensorFlow, PyTorch, and Hugging Face models) into a live, interactive web application with just a few lines of code. This open-source library eliminates the need for complex frontend development, handling all HTML, CSS, and JavaScript automatically. Developers define the function and specify inputs (e.g., 'text', 'image', 'slider') and outputs, then launch the interface locally, embed it in a notebook, or instantly generate a shareable public link. Gradio is widely adopted for quick prototyping, model demonstration, and deployment on platforms like Hugging Face Spaces, making complex models accessible to non-technical users for testing and feedback.

- Official technology site: https://www.gradio.app/
- Public AI Tinkerers demos and talks: 9
- Result page: 1 of 1

## Recent Public Talks and Demos

### [The Layer Between You and the Model](https://nashville.aitinkerers.org/talks/rsvp_vfOwHHG2DJ4)

When you send a message to an LLM, what happens before the model sees it? Something has to load context, decide what tools are available, shape what the model can and can't do. That's the harness. Everyone building agents has one, most people just don't call it that. This is a code walkthrough of a harness I've been building. I'll show what's actually in that layer: context loading, command constraints, persona switching, conversation branching. Not polished or a product pitch. Just the guts of what sits between me and the model. If you've been building something similar and didn't know the name for it, find me afterwards.

- 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_vfOwHHG2DJ4

### [Document Analytics with Open weight models for Sovereign Data](https://berlin.aitinkerers.org/talks/rsvp_5W86mvFdiH0)

get started with Document analytics in < 100 lines of code. PDF to Image -> VLM for OCR with Image -> Structured JSON Response -> Text to SQL with Tool Call -> Uses Qwen-3-VL-4B-Instruct

- Event context: AI Tinkerers Berlin Meetup - November 12th, 2025 — 2025-11-12 — Berlin
- Public talk page: https://berlin.aitinkerers.org/talks/rsvp_5W86mvFdiH0

### [AI Agents playing Texas Hold'em](https://seattle.aitinkerers.org/talks/rsvp_HCrImvLeCLw)

I'm working on a prototype that shows how a group of agents can play Texas Hold'em poker together. This is early work, but I'll be sharing progress, challenges encountered so far, and welcome any input/advice on frameworks or tools as I continue to develop this.

- Event context: October Meetup - Science Fair at Foundations — 2025-10-23 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_HCrImvLeCLw

### [Simon Says Prompts](https://san-diego.aitinkerers.org/talks/rsvp_n2qTEAdCvUc)

What if I told you a single embedding could make a language model say anything you want? In this lightning-fast demo, I’ll show you how to "hijack" an LLM’s output with just one vector. We’ll fire up the code, tweak an embedding live, and watch the model spit out whatever text we choose. An advanced trick with interesting implications: total control over LLMs!

- Event context: AI Tinkerers – San Diego Inaugural Meetup — 2025-09-12 — San Diego
- Public talk page: https://san-diego.aitinkerers.org/talks/rsvp_n2qTEAdCvUc

### [An MCP Server for the BBC micro:bit](https://seattle.aitinkerers.org/talks/rsvp_5Cvg0c_6Rew)

The BBC micro:bit is a small but powerful device, designing for learning and creating with technology. Over 7 million of them have been produced, used by over 30 million students in 70+ countries. In this science fair demo, I'll be presenting an MCP (Model Context Protocol) server I've developed for the micro:bit. Using MCP with the micro:bit opens up new possibilities for students - enabling them to interact with the device through conversational LLMs, going beyond what they can do today with the standard MicroPython implementation.

- Event context: Summer on the Lake - August Tinkerers Meetup! (Demos, Food, Friends) — 2025-08-28 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_5Cvg0c_6Rew

### [Anywhere MCP Mobile App](https://orange-county.aitinkerers.org/talks/rsvp_k3LbAklRqlY)

A comprehensive Model Context Protocol (MCP) platform that enables seamless AI tool integration across multiple devices. Anywhere MCP provides both real-time voice interactions through WebRTC streaming and REST API endpoints, allowing developers to connect AI agents with external tools and services anywhere - from web browsers to mobile apps to command-line interfaces. The roadmap for the platform will implement advanced agentic workflows including recursive planning, intelligent tool selection, and self-correction mechanisms, allowing AI agents to autonomously compose complex multi-step operations using external services like databases, APIs, file systems, and cloud resources. Code Examples: Real-time Tool Discovery & Validation ``` @log_mcp_operation("MCP tool registration") async def register_tools(self, session_id: str, tools: List[MCPTool]) -> RegistrationResult: # Dynamic schema validation with concurrent processing validation_service = get_validation_service() for tool in tools: validation_result = await validation_service.validate_tool(tool) if not validation_result.valid: # Self-healing: attempt parameter correction corrected_tool = await self._auto_correct_tool_parameters(tool) ``` Concurrent Session Management Architecture ``` class MCPSessionService: def __init__(self): self.max_tools_per_session = 50 self.session_tool_ttl = 24 * 60 * 60 # 24 hours @retry_on_redis_error(max_retries=3) async def execute_tool(self, tool_name: str, parameters: Dict[str, Any], session_id: str): # Handles 1000+ concurrent sessions with sub-200ms latency ``` Cross-Platform State Synchronization ``` // Vue.js composable maintaining real-time state across modalities export const useMCPSession = () => { const sessionTools = ref>(new Map()) const serverStatus = ref>(new Map()) // Real-time synchronization between WebSocket, REST, and mobile const syncToolState = async (sessionId: string) => { const response = await mcpApi.getSessionTools(sessionId) sessionTools.value.set(sessionId, response.tools) } ```

- Event context: AI Tinkerers - Orange County Meetup- Thursday July 31st 2025 — 2025-08-01 — Orange County
- Public talk page: https://orange-county.aitinkerers.org/talks/rsvp_k3LbAklRqlY

### [Building High-Performance Search Agents: Local Inference with DuckDuckGo and Google Search Integration](https://quito.aitinkerers.org/talks/rsvp_cMByAird1_Y)

Demostraré cómo crear agentes de búsqueda de alto rendimiento con inferencia LLM local usando llama-cpp-agents. Esta sesión de programación en vivo mostrará cómo implementar integraciones de búsqueda de DuckDuckGo y Google, optimizar la gestión de memoria para ventanas de contexto grandes (32 000 tokens) y lograr una latencia un 60 % menor que las alternativas en la nube. Verá técnicas prácticas para el uso eficiente de tokens, la aceleración de GPU con flash attention y la gestión de citas de fuentes. La demostración incluirá métricas de rendimiento en tiempo real que comparan ambos motores de búsqueda en los modelos open source

- Event context: AI Tinkerers - Quito Primer Meetup (Abril) — 2025-04-24 — Quito
- Public talk page: https://quito.aitinkerers.org/talks/rsvp_cMByAird1_Y

### [A2A Deep dive](https://dublin.aitinkerers.org/talks/rsvp_PUfCKEuGrfQ)

Agent To Agent (A2A) is a newly proposed protocol by google that aims at standardizing how Agents communicate with other vendors, this is particularly useful in the case of 3rd party AI Agents. In this demo, I'll walk through some of the key design principles of A2A, by walking through code and highlighting how the principles are/can be implemented. We'll step through what a client and server for A2A should look like and how communication between the two is facilitated by a task manager.

- Event context: AI Tinkerers - Dublin Event (April) — 2025-04-24 — Dublin
- Public talk page: https://dublin.aitinkerers.org/talks/rsvp_PUfCKEuGrfQ

### [Long term and short term memory](https://la.aitinkerers.org/talks/rsvp_yDt1OftJcCQ)

Wrote a quick gradio script to use chromadb to act as a long term memory of past conversations. Still a wip, but it works and would like to share it with others. The tool mimics anything-llm but in 250 lines of code. I intend on having web search functionality built in. Basically swap in any openai compatible api endpoint (use text-generation-webui) and you're golden. For the demo, I plan on using openai's endpoint.

- Event context: March 19th - LA AI Tinkerers Meetup & Demos — 2024-03-20 — Los Angeles
- Public talk page: https://la.aitinkerers.org/talks/rsvp_yDt1OftJcCQ

## Related Technologies

- [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) — 664 public demos
- [FastAPI](https://aitinkerers.org/technologies/fastapi) ([Markdown](https://aitinkerers.org/technologies/fastapi.md)) — 181 public demos
- [LangChain](https://aitinkerers.org/technologies/langchain) ([Markdown](https://aitinkerers.org/technologies/langchain.md)) — 446 public demos
- [llama](https://aitinkerers.org/technologies/llama) ([Markdown](https://aitinkerers.org/technologies/llama.md)) — 48 public demos
- [OpenAI Agent SDK](https://aitinkerers.org/technologies/openai-agent-sdk) ([Markdown](https://aitinkerers.org/technologies/openai-agent-sdk.md)) — 4 public demos
- [Anthropic](https://aitinkerers.org/technologies/anthropic) ([Markdown](https://aitinkerers.org/technologies/anthropic.md)) — 36 public demos
- [BBC micro:bit](https://aitinkerers.org/technologies/bbc-micro-bit) ([Markdown](https://aitinkerers.org/technologies/bbc-micro-bit.md)) — 1 public demo
- [Chroma](https://aitinkerers.org/technologies/chroma) ([Markdown](https://aitinkerers.org/technologies/chroma.md)) — 8 public demos
- [crew](https://aitinkerers.org/technologies/crew) ([Markdown](https://aitinkerers.org/technologies/crew.md)) — 1 public demo
- [Flash Attention](https://aitinkerers.org/technologies/flash-attention) ([Markdown](https://aitinkerers.org/technologies/flash-attention.md)) — 1 public demo
- [GGUF](https://aitinkerers.org/technologies/gguf) ([Markdown](https://aitinkerers.org/technologies/gguf.md)) — 5 public demos
- [Groq](https://aitinkerers.org/technologies/groq) ([Markdown](https://aitinkerers.org/technologies/groq.md)) — 23 public demos
- [HTTP/1](https://aitinkerers.org/technologies/http-1) ([Markdown](https://aitinkerers.org/technologies/http-1.md)) — 1 public demo
- [JSON-RPC](https://aitinkerers.org/technologies/json-rpc) ([Markdown](https://aitinkerers.org/technologies/json-rpc.md)) — 2 public demos
- [llama-cpp-agents](https://aitinkerers.org/technologies/llama-cpp-agents) ([Markdown](https://aitinkerers.org/technologies/llama-cpp-agents.md)) — 1 public demo
- [MCP](https://aitinkerers.org/technologies/mcp) ([Markdown](https://aitinkerers.org/technologies/mcp.md)) — 129 public demos
- [OpenAI](https://aitinkerers.org/technologies/openai) ([Markdown](https://aitinkerers.org/technologies/openai.md)) — 112 public demos
- [OpenAI API](https://aitinkerers.org/technologies/openai-api) ([Markdown](https://aitinkerers.org/technologies/openai-api.md)) — 520 public demos