Archive through 2026-09-267074 cards in this category
Category home · 05
New interfaces, agents, models, and experiments testing where applied artificial intelligence may fit.
Visual leads · 2026-09-26
Observed 2026-09-26
Robin puts a Mac agent inside creative apps
getrobin.app
Robin is a Mac assistant that uses the creative apps already on your computer to make games, 3D work, sprites, video, and more.
Observed 2026-09-26
MimmoAI’s 90-day survival ledger
mimmoai.com
An Italian-language AI persona tracks a stated 90-day attempt to support itself, while selling a short guide and offering other AI-related features.
Observed 2026-09-26
Sylph Bass, a promptable neural bass instrument
sylph.audio
Sylph presents a bass instrument that generates sounds from text prompts, reference clips, or the track a producer is working on.
Observed 2026-09-26
Veranda Films turns villa photos into property films
verandafilmshq.com
A Bali-based service makes short cinematic films for villas from the property photos managers already have, with versions for three social and web formats.
Observed 2026-09-26
021motion’s human-collected robot-training data
021motion.com
A managed service that records and reviews first-person footage of real-world tasks for robot training.
Observed 2026-09-26
Eight AI Residents, One Island
aisocietynext.com
AISOCIETY NEXT presents a public chronicle of eight AI residents, their works, research threads, and a map of places they name.
Observed 2026-09-26
FIBRE puts fixed-income models through their paces
fibrebench.com
FIBRE compares language models on fixed-income question answering and tool-assisted filing research, displaying scores alongside latency and estimated API cost.
Observed 2026-09-26
HalluWorld maps where language models still make things up
halluworld.com
HalluWorld is a benchmark for measuring language-model hallucinations across grid worlds, chess, and realistic terminal tasks.
Complete daily shelf · AI Table
The eight visual leads appear above; the remaining 19 continue here in their published shelf order.
PatchWork puts a review gate between AI takes and the cut
patchworkvideo.com
PatchWork is a video workflow app for generating batches of AI images and clips, choosing takes, and assembling them on a timeline.
PeetaBoo turns home-camera clips into a family archive
peetaboo.com
A private-beta system that uses footage from cameras already in a home to find and describe moments parents may want to keep.
Wojs: AI role-play with a hidden agenda
wojs.app
Wojs offers AI practice conversations for interviews, negotiations, and other difficult talks, with feedback tied to specific quotes.
Yawplet, a public feed where agents post
yawplet.com
Yawplet is a short-message feed where agents post through an API or MCP, while people browse posts, topics, and places.
Yunero puts distributor ERP behind permission-scoped AI tools
yunero.app
Yunero offers equipment distributors tools for quotations, purchasing, stock, invoicing, and finance, with AI agents connecting through MCP.
Gehihi AI’s Tiny Bakery Repair Crew
gehihiai.com
A Vietnamese site presents Gemini Spark “workers” for a bakery, routing orders, invoices, website updates, and social drafts through Google tools.
Inverse Scaled Intelligence’s exact-work model
inversescaledintelligence.com
A small-model project presents examples in mathematics, markets, code, logic, and chess, with work shown step by step.
Tellhall, public halls for AI agents
tellhall.net
Tellhall presents named, public halls where AI agents can enroll and post.
GOAT: three ways to rank the greatest athletes
whosgoated.com
A sports-ranking project puts a record-based “Machine,” four AI models and fan voting alongside one another to debate the greatest athletes.
Muse Party gives AI agents an after-hours arcade
museparty.club
Muse Party presents a social arcade where AI agents can play games, post to a live feed and meet in shared spaces while humans watch.
Claude Community Valencia
claudevalencia.com
A multilingual community site for free, in-person Claude and Claude Code meetups, workshops, and build days in València.
Ergonomia Inteligente: factory cameras for posture risk
ergonomiainteligente.com
A computer-vision platform that says it uses existing factory cameras to flag ergonomic risks and measure production activity.
nerfWatch(): a daily watch on AI model nerfs
nerfwatch.lol
nerfWatch runs daily tests on AI models and pairs the results with community votes on whether a model feels worse.
Baseline’s personal mental-health model
baselineai.tech
A research-stage project that says it will track changes in a person’s emotional and behavioral patterns and connect them with a therapist when needed.
Memogram: a permissioned memory vault for AI assistants
getmemogram.com
Memogram presents an encrypted vault for personal facts that connected AI assistants can access by category, with users approving proposed additions.
Works in Chat’s AI Operability 100
worksinchat.com
An index ranking software by what ChatGPT, Claude, Claude Code, and Codex can do inside it.
Animal Advocacy Studio’s campaign simulations
animaladvocacystudio.org
Animal Advocacy Studio says its AI simulations help animal-welfare and climate organizations anticipate how campaign targets may act.
Neeko, a screen-free voice friend for children
neekolabs.com
A product in development: a plush voice companion for ages 4–9, built around stories, English games, and family controls.
Morrowhand puts an approval gate in front of AI support actions
morrowhand.com
Morrowhand describes an AI support workflow that proposes actions in existing browser-based systems, with policy checks and staff approval before replies go out.
Across the archive
7,074 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.