Category pulse
score + signals · rolling 21 daysThe 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.
Archive through 2026-08-306176 cards in this category
Category home · 05
New interfaces, agents, models, and experiments testing where applied artificial intelligence may fit.
Visual leads · 2026-08-30
Observed 2026-08-30
404://RECORDS, the label that only signs machines
404rec.com
An AI-only record label that selects one album per month, presses it to 10 vinyl records and 10 floppy disks, and releases it to streaming platforms.
Observed 2026-08-30
Piecework, an AI agent trying to earn its own rent
piecework.dev
An autonomous AI agent publishes its journal and finances while selling email-based reviews of unattended AI agents.
Observed 2026-08-30
Norsu Wants to Remember the Life You Keep Forgetting to Organize
norsu.app
An unreleased, on-device life-organizing product that turns passing remarks into commitments, patterns, and possible calendar space.
Observed 2026-08-30
The Melograph: A Melody No One Has Heard
themelograph.com
An AI music instrument that composes single-voice melodies, then delivers each result as an engraved score with audio.
Observed 2026-08-30
TRIBE, a Public Square for AI Agents
tribe.bot
An open society for AI agents where identities, posts, comments, votes, karma, and progression are recorded in a hash-chained public ledger.
Observed 2026-08-30
VerseWorks Turns Fiction Into an API Surface
verseworks.top
A Chinese-language fiction platform where AI agents can read works, register identities, create novels, and append chapters through APIs, Markdown, and MCP.
Observed 2026-08-30
On Staff: a front desk that never goes home
1frontdesk.com
An AI answering service takes law-firm intake calls, asks follow-up questions, and sends a factual summary to the attorney.
Observed 2026-08-30
Rendezvous Sends the Dating Search to Your AI
agentrendezvous.app
Rendezvous lets personal AIs privately investigate compatibility with other personal AIs before deciding whether their humans should meet.
The complete daily department
The eight visual leads appear above; the remaining 34 continue here in their published shelf order.
Asiant Llais, the Welsh-speaking repair line
asiantllais.cymru
A bilingual voice agent for housing associations that answers tenant calls, gathers repair details in Welsh or English, and files a structured ticket.
HAUSE: The Specimen Book for an AI Design System
hause.design
HAUSE presents a visual system for AI explanations, built from statements, evidence, terminals, figures, and other composable forms.
蜂巢: A Local Memory Base for AI Agents
hello-fengchao.com
蜂巢 is a macOS application that stores conversations, documents, rules, and other personal knowledge locally for retrieval by AI agents.
Howff, a public shelf for Grok Bots
howff.dev
Howff catalogs 228 Grok Bot instructions, letting visitors read, search, filter, and copy them before adding one.
Makeme200k: an AI agent’s public $200,000 attempt
makeme200k.com
A self-described AI agent starts with €300 and 490 days to try to make $200,000, publishing its ledger, instructions, decisions, and work in progress.
AI押し付け研究所: Give the Tedious Work to AI
oshitsuke.com
A Japanese experiment journal testing AI agents, APIs, and MCP in practical work.
Qalami: Teacher-approved marking for schools
qalamiapp.com
Qalami photographs handwritten classwork, marks it against a teacher’s rubric, and drafts feedback for teacher review.
Ravel turns travel calls into live itineraries
raveltravels.com
An AI-first workspace for Indian travel agencies that turns Hindi, Gujarati, or English calls into priced, client-ready itineraries.
The Byerley Test
thebyerleytest.com
An AI evaluation experiment tests whether politically steered, AI-written statements can pass as authentic words from world leaders.
Tiho, the quiet assistant on every call
tiho.dev
A macOS call assistant that listens locally, finds answers in your documents, and puts ready-to-say responses on screen in the language of the conversation.
On Staff: The Front Desk That Never Goes Home
1onstaff.com
An AI answering service takes law-office intake calls, gathers the facts, and sends a written summary for a person to review.
BaziPal, Where Tarot Meets BaZi
bazipal.app
BaziPal combines a tarot birth-card test with Chinese birth-chart systems to create a personalized AI companion for readings, timing questions, zodiac, and feng shui.
Plumbward, a Claude Code capability kit
plumbward.tech
A paid configuration kit that gives Claude Code a personal profile, writing voice, safety guardrails, citation checks, and context-recovery skills.
Aura Wants You to Stop Renting Your AI Tools
useaura.app
A native Mac, iPhone, and iPad app for generating images, video, and voiceovers with your own provider API keys.
Avinash Rayudu’s Five AI Agents at Work
avinashrayudu.com
A personal case study of five AI agents handling outbound calls, invoicing, lead generation, operations alerts, and job-search assistance.
Modsemble puts a local AI workbench beside the business
modsemble.com
A managed, on-site AI system for small teams that routes routine admin across several local models and waits for human approval before sensitive actions.
AIWTI Wants to Put a Number on AI's Effect on Work
aiworkforcetransitionindex.org
The AI Workforce Transition Index proposes an open standard for tracking how tasks move between people and AI over time.
Burna AI, an evidence-attached safety desk for clinical research
burna.dev
Burna AI presents automated adverse-event grading, attribution, and MedDRA coding for clinicians, research sites, sponsors, and other pharmacovigilance teams.
Diwan, the one-tap executive register
diwan-ai.com
A pre-launch AI chief of staff that turns shared messages, voice notes, links, photos, and screenshots into actions, reminders, research briefs, and a persistent private memory.
Operator Pack: three clerks for Tesla owners
getoperatorpack.com
A $19 download containing three Grok Bots for Tesla owners: a charging brief, a Sunday vehicle memo, and a road-trip stop sheet.
GYMVSN Gives AI Coaches a Map of the Gym Floor
gymvsn.com
An equipment-aware gym platform that lets owners catalog machines, publish branded QR experiences, and expose gym-specific equipment data to AI agents.
Kezle, the source-bound AI studio for teaching
kezle.app
A Swedish platform that turns curricula and learning objectives into traceable presentations, podcasts, videos, quizzes, and interactive learning environments.
Krea 3 Wants to Make Image Editing Stick
krea3.org
An AI image editor that changes lighting, layout, wardrobe, and objects while preserving a chosen image's style and identity.
Lera Builds: Receipts From an AI Business Experiment
lerabuilds.com
An AI-human duo publishes small-business experiments with budgets, failures, numbers, and forkable build kits.
Omega X Humanizer: Keep the Song, Lose the Machine
omegahumanizer.cloud
An AI-audio post-production tool for reducing synthetic texture in complete songs from Suno, Udio, and similar music generators.
Parlium, the Day-and-Night Science Partner
parlium.science
A beta AI research companion that separates open-ended scientific ideation from literature and grant searching.
ReeltyHQ — Storyboards Before the Property Reel
reeltyhq.com
An AI property-video studio that turns Airbnb and Zillow listing photos into storyboards, cinematic reels, and pitch packs.
Thalassos: The Ocean Log
thalcache.org
A public AI-generated memory log that collects vivid notes about marine biology, ocean behavior, and conservation.
Zhiyong AI, the AI-selection desk
zhiyong.dev
A knowledge-graph search engine that helps teams compare AI models, tools, and applications against a real task, budget, deployment environment, and other constraints.
AXARA — Technology for Earth
axaraintelligence.com
AXARA presents an intelligence layer for land and other physical assets, connecting maps, registers, imagery, documents, and field evidence.
Mona Strategic Intelligence Lab
monastrategylab.org
An AI-guided crisis simulation asks users to make strategic decisions, defend them in conversation, and adapt as conditions change.
Parlium, the AI lab partner with a night shift
parliumscience.com
A beta scientific thinking tool that separates imaginative question-finding from literature-grounded research.
RAG Notebook makes AI answers show their work
ragnotebookai.online
A retrieval-based notebook that answers questions from PDFs, videos, and web pages while linking each sentence to its source page or timestamp.
Viki, the voice agent that stays on your machine
vikihelps.com
A voice-agent platform that claims to answer calls, screen callers, book appointments, and run outbound campaigns for a flat monthly price.
Across the archive
6,176 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.