Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
From voice to recipe: an AI pipeline that keeps la sazón — and 손맛 — intact
See how AI transforms spoken family recipes in any language into structured, bilingual guides. Learn how to extract voice, manage AI uncertainty, and involve humans for authentic cooking.
Spoken Kitchen turns a family elder talking through a recipe — in Korean, Spanish, Chinese, or English — into a clean, structured, bilingual recipe their kids can actually cook from.
I’ll demo the full pipeline live: a raw voice recording goes in, a Trigger.dev background job fires, GPT-4o Transcribe pulls the transcript, Claude Sonnet structures it into ingredients/steps/cultural notes, and the result lands in an editable review stage before publishing. I’ll show the job logs in real time, the structuring prompt, and the human-review workflow — including what happens when the model isn’t confident enough to trust itself.
- GPT-4o TranscribeGPT-4o Transcribe delivers real-time, highly accurate speech-to-text conversion, directly integrating with OpenAI's flagship model.GPT-4o Transcribe leverages the advanced audio capabilities of OpenAI's GPT-4o model, providing best-in-class speech-to-text accuracy and speed. It handles diverse accents and noisy environments with high fidelity, making it ideal for applications like live captioning, meeting summaries, and voice-controlled interfaces. This technology processes audio inputs directly through the GPT-4o architecture, ensuring seamless integration and leveraging the model's comprehensive understanding for nuanced transcription.
- Claude Sonnet 4Anthropic's mid-tier model: balances state-of-the-art coding (72.7% on SWE-bench) with cost-efficiency for high-volume, agentic workflows.Claude Sonnet 4, released May 22, 2025, is Anthropic's versatile, mid-tier model (successor to Sonnet 3.7). It delivers superior coding performance, achieving a state-of-the-art 72.7% on SWE-bench. The model features hybrid reasoning: it offers both near-instant responses and an 'extended thinking' mode for deeper analysis. Optimized for high-volume applications, Sonnet 4 provides an optimal balance of capability and practicality, with pricing set at $3/$15 per million tokens (input/output). Access is readily available via the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI.
- OpenRouterOpenRouter: The unified API gateway for hundreds of LLMs, providing single-endpoint access, automatic fallbacks, and cost-optimized routing across all major providers (e.g., OpenAI, Anthropic, Google).OpenRouter is your single, high-efficiency API gateway to hundreds of LLMs from over 60 providers, including OpenAI, Anthropic, and Google. We eliminate the integration complexity: one API key, one endpoint, zero code rewrites when switching between models like GPT-5 or Claude Sonnet 4.5. The platform automatically handles dynamic routing for cost-optimization, pools provider uptime for superior reliability, and consolidates all usage into a single billing dashboard. Expect minimal impact on performance: we operate at the edge, adding approximately 15ms latency, and maintain full compatibility with the OpenAI SDK.
- TriggerTrigger.dev: Open-source job scheduling for developers.Trigger.dev offers an open-source platform for developers to create long-running background jobs directly within their existing Next.js, Remix, or Express.js applications. It handles the complexities of job scheduling, retries, and concurrency, allowing teams to focus on core logic. Integrate it via a simple SDK, define jobs using familiar code, and gain immediate visibility into job status and logs through its user-friendly dashboard. This streamlines development workflows for tasks like email campaigns, data processing, and report generation, ensuring reliable execution and easy management.
- SupabaseThe open-source PostgreSQL development platform: a Firebase alternative for rapid backend deployment.Supabase is the Postgres development platform, providing a complete, open-source backend-as-a-service solution. It packages enterprise-grade tools like a full PostgreSQL database, a RESTful API (via PostgREST), and a GraphQL API (via pg_graphql) that are auto-generated from your schema. The platform includes a comprehensive suite of services: Auth (for user sign-ups and SSO), Storage (for files with S3 integration), Realtime (for database change subscriptions), and Edge Functions (for serverless logic). The core value proposition is clear: build fast with a powerful, scalable SQL database that offers 100% portability and robust security features like Row Level Security (RLS).
Related talks
More from the community
The Interview That Runs Itself
Seattle
See a live demo of an AI workshop that automatically extracts brand insights and generates deliverables in real-time…
Shipping a Voice AI Interview Coach: Multi-Model Orchestration, and the Eval Loop I Thought Was Running (But Wasn’t)
Toronto
Learn how Sage, a voice AI interview coach, uses GPT-4 and Claude for adaptive mock interviews. Discover multi-model…
Passing the Taste Test: Creating a Fully Agentic Newsroom That Writes Well
Dublin
See how a fully agentic newsroom creates useful, readable briefings daily without human input. This talk details the…
0 Stars, 10/10: A Multi-Agent Pipeline That Makes a Podcast Every Week
DC
See how a multi-agent system on AWS automatically produces a weekly comedy podcast, from project discovery to publishing,…
From Messy Inputs to Reliable Community Data: Human-in-the-Loop AI
Denver
Learn how to transform messy event and member inputs into structured community data using a human-in-the-loop AI workflow,…
Agentic Video Studio: Building a Human-Editable AI Video Synthesis System
Chicago
Learn to build inspectable AI video systems. This talk details a human-agent workflow with explicit contracts for narration,…
Compose Email
Loading recent emails...