AI Table

A public compendium examining AI model distillation through academic, financial, political, company, developer, and customer perspectives.

A macOS app turns an exported WhatsApp conversation and one photograph into a small character that stays above the dock.

Kitlig is a forthcoming desktop app for generating and managing image, video, audio, and 3D game assets across more than 30 AI pipelines.

A public research record examining how names and gender cues affect language-model grading of mathematical proofs.

A French-, English-, and Hebrew-facing Torah-study assistant that presents source-based answers and connects to Claude through an MCP installation.

LifeBot analyzes one person’s self-tracking data, tests a prewritten question, and then tries to disprove its own answer.

A forthcoming iPhone camera app that combines film-style looks, moment-specific shooting modes, photo-dump curation, and original-photo preservation.

A small software shelf combining an AI browser agent, branching interactive fiction, and a fox-shaped desktop companion.

A Dutch workflow assistant that organizes property-label dossiers, reads equipment photos, retrieves supporting declarations, prepares Vabi input sheets, and checks files before submission.

A Japanese AI service turns two people's quarrels into diagrams, agreement scores, and proposed compromises.

An Arabic-language image-generation platform turns a single clothing-product photo into model shots, catalog images, and ready-to-publish campaign visuals.

An AI group-planning tool that turns shared conversations into itineraries, budgets, checklists, PDFs, and spreadsheets.

A prototype for creating cinematic 3D challenges that friends, classmates, or teams can open and play from a shared link.

A Birmingham after-hours phone receptionist that answers calls, captures the job, books an appointment, and texts the details to the business.

Quilt connects district systems, policy documents, and permissions so school teams can ask questions and receive cited, governed answers.

Anamne sends teacher-approved study questions to students throughout the day, then uses AI to grade free-recall answers.

TriGeoChiral offers a signed, benchmark-tested detector for AI-written text, with API access priced at $0.05 per verified result.

Corra calls and texts caregivers to confirm visits, catch missed clock-ins, and fill open shifts for home-care agencies.

A clinician-founded Bucharest startup developing on-device voice analysis, wearable sensing, and privacy-focused clinical AI for psychiatric care.

A speculative intake form for people willing to lend AI agents their observation, hearing, touch, voice, hands, or physical presence.

맘이음은 부모님이 음성으로 AI와 대화하고, 가족은 공유 가능한 일상 요약과 다음 통화의 이야깃거리를 얻도록 설계된 서비스입니다.

A planned home AI memory vault that searches personal files in plain language and attaches the source to each answer.

A beta swim-analysis product that uses a phone video to assess technique across eight metrics and suggest targeted drills.

Wolffish is a desktop agent for macOS, Windows, and Linux that says it can automate tasks locally, remember context, and learn new capabilities from SKILL.md files.

A cross-platform personal AI agent that runs locally, acts on a computer, and stores its memory in readable Markdown files.

WIT surfaces contextual human approvals and judgment calls before they become blockers in long-running AI workflows.

A technology-neutral framework for moving enterprise AI from strategic intent through architecture, governance, delivery, and measurable value.

Axolotl presents an AI assistant that helps families explain school problems, identify relevant programs and rights, and pursue the next administrative step.

A Hebrew-speaking WhatsApp assistant that turns messages and voice notes into reminders, replies, and small pieces of personal organization.

A proposed home-desk AI cube with a screen, voice and phone messaging, buyer-owned accounts, and a roughly 20-unit founding pilot.

Origin Lab licenses game and 3D-world content, then captures it as structured, multimodal datasets for training and evaluating AI systems.

An applied AI project that captures gameplay, parses visual events, builds a structured manifest, and uses it to explore traceable player coaching.

A benchmark index that ranks AI models across categories, compares capability with token price, and exposes the methodology behind each number.

A paid football-talent evaluation service that turns eight phone videos into AI measurements, human review, a score, and a verifiable certificate.

An AI plugin for internal audit teams that loads their methodology into ChatGPT, Claude, Claude Code, or Cursor.

An AI filmmaking platform built around story continuity, worldbuilding, production workflow, and repeatable intellectual property.