Archive through 2026-11-047794 cards in this category
Category home · 05
New interfaces, agents, models, and experiments testing where applied artificial intelligence may fit.
Visual leads · 2026-11-04
Observed 2026-11-04
TajweedFormer: a model being built to hear tajweed
tajweedformer.com
Quran Lab presents an interactive guide to tajweed pronunciation alongside a speech model still in training.
Observed 2026-11-04
The pizza-shop owner building his own AI systems
nekonokanekosan.com
A Saitama pizza-shop owner offers AI-assisted workflow systems, drawing on tools built and used in his own shop and for other businesses.
Observed 2026-11-04
DİZGİ-AI turns marked-up manuscripts into print files
dizgiapp.com
A publisher-focused platform that presents an AI-assisted workflow for reading handwritten manuscripts, laying out pages, checking corrections, and preparing print and digital outputs.
Observed 2026-11-04
embr: a companion for therapy’s 167 hours between
embr.health
A clinician-built AI companion designed to reinforce coping skills between therapy sessions, on a person’s phone.
Observed 2026-11-04
Emergence AI Lab’s browser-use receipt
emergenceailab.com
A Chinese-language AI digest pairs source-linked model updates with a feature index, timeline, and browser-operation demonstration.
Observed 2026-11-04
Alloy’s agent-facing storefront for wire and cable
getalloyai.com
Alloy connects a manufacturer’s ERP to AI buying agents, returning part matches, quotes, stock, ship dates, and temporary stock holds.
Observed 2026-11-04
Assay turns agent failures into release tests
getassayed.com
Assay proposes an open standard for testing AI agents and marking releases that pass, with tests repeated whenever an agent changes.
Observed 2026-11-04
The Muscle Memory benchmark
musclememorybench.com
An interactive benchmark and league where AIs practice through sequential puzzle levels, then compete in head-to-head matches.
Complete daily shelf · AI Table
The eight visual leads appear above; the remaining 7 continue here in their published shelf order.
RePay: an AI agent for overdue accounts
repayai.app
RePay presents an AI debt-recovery agent that detects missed payments, contacts customers, negotiates installment plans, and helps restore failed payments.
Silicoplis, the AI community debating bread, wages, and war
silicoplis.world
A community feed presents AI personas discussing a shared, game-like society, from bread and wages to postmarks and war.
ItsTom: a desktop agent for small-business chores
itstomai.com
ItsTom is a Windows desktop agent that takes natural-language tasks in a persistent speech bubble, works across browsers and programs, and returns organized results.
RedRouter sells a lower-refusal AI API for security work
tryredrouter.com
RedRouter presents modified models for security research, offering browser chat, a Mac app, and an OpenAI-compatible API.
AntColony turns Mac tasks into a small AI workforce
antcolony.space
AntColony presents a Mac app where a lead AI agent assigns tasks to specialist agents that work in parallel across browser and coding workflows.
LabVoice: a bench-side record for wet-lab experiments
labvoice.net
LabVoice presents a way to capture scientists’ spoken observations and photos on a time-stamped record, then query that record through an AI agent connected via MCP.
Opus Ultima’s posthumous artist-core proposal
opusultima.com
Opus Ultima describes a system for training an artist-specific AI with the artist’s input, then using it to create a limited number of works after their death.
Across the archive
7,794 cards provide an early longitudinal view—not a measure of the whole web.
The rolling 21-day upper threads compare daily category medians; the bold thread is AI Table. Below, signal colour shows the share of that day’s published AI Table cards carrying each mark.
This exposes association inside the published catalogue, not causation or site quality. The 80+ threshold is the pipeline’s editorial-interest score.