Interactive demonstrations of Large Language Model applications — explore real-world AI in action.
Try a different search term.
Gallery of AI-assisted data journalism cards for Indian newsroom stories, with MOSPI, UN, and survey-driven explainers plus verification workflow support.
AI-assisted data analysis workspace that generates hypotheses, tests them statistically against uploaded datasets, and organizes findings into explainable artifacts.
Interactive benchmark report comparing multi-model speech transcription quality, latency, evaluator behavior, and model limitations across audio variations.
Interactive demo and API guide for a modern D3 word-cloud layout with async placement, incremental streaming, bitmap collision, and SVG rendering helpers.
Collaborative editing prototype where humans and LLM agents work on shared documents with live activity, manual instructions, and edit explanations.
Exploration report showing LeWM world-model demos for visual planning, physics anomaly detection, and latent-space probing on robot-control tasks.
Audio similarity dashboard that trims tracks, separates stems, generates Gemini embeddings, and lets users explore song clusters, comparisons, and heatmaps.
UMAP-based research intelligence demo that maps SNHU publications against OpenAlex data and highlights field evolution and strategy opportunities.
Synthetic education analytics portal with data stories, dashboards, leaderboards, intervention triage views, and an executive briefing deck.
Data story benchmarking local and LLM-based OCR models on financial documents with multi-judge scoring, category breakdowns, and cost-quality analysis.
QSR pricing simulator for exploring menu price changes, customer segments, channel migration, churn elasticity, revenue, orders, and contribution margin.
Interactive operating model for testing Taco Bell item-price changes against demand, channel mix, sales, and margin before rolling them across the menu.
Experiment report showing how Claude Code and Lean 4 were used to machine-check 22 theorems, including a formal proof of IMO 1977 Problem 5.
Build-log narrative of wiring Claude Desktop to MATLAB through MCP, debugging the setup, and then using the bridge to run five progressively richer simulations.
Geospatial viewer that overlays historical lake imagery with SAM3 water segmentation so users can compare temporal change and estimated surface area.
Projects CSV rows into a similarity map with UMAP so users can browse large tabular datasets as clusters, grids, lists, or tables.
Arcade survival game where players dodge spreading lava, chain powerups, and compete for leaderboard scores in short browser runs.
Transforms 150 Arijit Singh tracks into animated waveforms that can be filtered by mood while exposing BPM, energy, and preview clips.
Maps chart screenshots from major publishers into a UMAP scatterplot so readers can compare how newsroom visual styles cluster by similarity.
Turns microphone or uploaded audio into an interactive galaxy whose motion reacts live to sound amplitude and frequency.
Generates company research, slide decks, narration, and exportable briefing videos in one browser workflow.
Encodes insurance contracts as logic rules and animates claim-verification traces so users can inspect how each coverage decision is reached.
Analyzes district healthcare datasets and turns anomalies into prioritized action plans and email drafts for government administrators.
Explores six public datasets with AI-ranked anomalies so readers can jump from raw records to the strangest patterns hiding in science and government data.
Shows how AI coding agents can emulate Stripe PaymentIntents, generate doc-traceable tests, and compare emulator behavior against live API responses.
Technical walkthrough of an LLM-to-robot pipeline that converts natural-language prompts into live ROS2 commands across Gazebo and MuJoCo simulators.
Script-backed casefiles for public-sector datasets that surface impossible totals, reporting breaks, and suspicious patterns with reproducible evidence.
Analyzes audience, inventory, and booking data to structure ad campaigns with better yield and fewer compliance surprises.
Walks through an AI-led Blender design session so teams can see where conversational 3D design works, fails, and improves through iteration.
Shows where loan approvals are getting stuck so operations teams can remove bottlenecks before deals miss deadlines.
Helps marketing teams plan and launch cross-channel campaigns from one workflow, reducing handoffs and missed compliance steps.
Helps trial teams catch risky or inconsistent clinical records early and route only tricky cases to human reviewers, saving review time.
Gives leaders one view of KPI health across teams so they can spot problems early and prioritize fixes.
Turns a messy API spec into test cases so QA teams can catch data-pipeline failures earlier and cut manual testing effort.
An interactive walkthrough of migrating 100 SQL Server scripts to MySQL with LLM-assisted conversion, verification, and business-impact simulation.
Converts screen recordings into runnable Playwright tests, helping teams reproduce bugs and automate regression checks faster.
Executive portal for interactive AI demos in enrollment triage, credit checking, and financial aid support for schools and universities.
Evaluates mixed student evidence such as transcripts, military records, and certifications to quickly estimate transfer credits with AI-guided extraction.
Enrollment operations dashboard that classifies inbound inquiries, drafts responses, routes escalations, and tracks response-time and coverage metrics.
Lets architects compare generated 3D plans with original 2D layouts, making it easier to pick models that preserve room accuracy.
Helps sales teams evaluate deal options faster by surfacing pricing and compliance risks and recommending safer, higher-margin scenarios.
Converts plain-English workflow descriptions into interactive metro-style pipeline maps with tool-level detail views.
Interactive gallery of distinct LLM-generated visual filters applied across chart, comic, and photo reference images.
Florida private school compliance checker that maps school documents to state and federal regulations and summarizes checklist-level compliance.
Summarizes parent feedback into clear pain points and school-level heatmaps so leadership can prioritize the highest-impact improvements.
Scrollytelling forensic report on NFL draft medical risk patterns, highlighting hidden pathologies and compensation signals across prospects.
Generates interactive map-based biographies from a person or storyline prompt, with shareable story views and JSON exports.
How 3,200+ researchers across 113 countries view AI, integrity, collaboration, mobility, and impact — from Elsevier's Confidence in Research report.
We analyzed 68,658 Retraction Watch records to find the biggest patterns, surprising signals, and practical actions. An early-warning publisher dashboard for where trust, reputation, and legal risk can spike.
Real-time dual-audio transcription assistant for support calls with live AI Q&A, summaries, and response drafting.
Editorial-comic generator and gallery that turns current headlines into satirical comic cards with filtering by date and topic.
Shows which SQL queries consume the most effort and runtime, helping data teams optimize critical reports and reduce repetitive ad-hoc work.
Scrollytelling data story comparing Nutella product formulations across countries to reveal regional differences in sugar content and market adaptation.
Transforms standard charts into handcrafted visual styles such as embroidery, pixel art, blueprint, and oil-paint aesthetics.
Compares two text-grouping methods on sample datasets so analysts can choose the approach that clusters their documents better.
Compares vision models on comic transcription quality so teams can pick the one that best preserves dialogue, speakers, and panel structure.
Benchmarks text-to-SVG models so designers can choose generators that produce cleaner, more usable artwork with less manual fixing.
Geospatial narrative explorer that presents curated urban-change stories with map-driven evidence and regional context.
Tracks urban change from satellite imagery so planners can see where growth, density, and green cover are shifting over time.
AI-powered fact-checking analysis of NCERT Class 12 history textbooks, verifying 45+ claims and identifying accuracy issues using automated verification against academic databases and archaeological records.
Turns noisy prediction-market activity into digestible trends, helping teams quickly see where consensus is strengthening or drifting.
Understand how price changes impact subscriber behavior, revenue, and growth for streaming services. Make data-driven pricing decisions with confidence through elasticity analysis.
An investigation into what happens when artificial intelligence reviews scientific papers — and what goes hilariously (and seriously) wrong.
Journey through Python's None-handling underworld by filing identical type errors in SQLAlchemy, Apache Arrow, and scikit-learn to reveal patterns in open-source maintenance and software evolution.
Scenario-based communication simulator that uses AI to pressure-test negotiation, crisis response, and persuasive speaking.
Compares image outputs from different AI models with visual error maps, helping teams choose the model closest to a target reference.
Generate hand-drawn SVG animations using AI from natural language prompts. Customize roughness, stroke width, timing, and colors to create beautiful animated drawings with realistic hand-drawn effects.
Visualize how language models predict the next token by exploring the 'multiverse' of paths the Gemma model considers. Interactive token trees reveal how probability shapes AI creativity and decision-making.
Multiple visual storytelling perspectives on open source community data. Explore contributor ecosystems through narrative journalism, interactive scrollytelling, artistic D3.js visualizations, and analytical insights.
How humans label emotions in music—400 tracks across 4 genres rated by multiple annotators, revealing which emotions spark the most disagreement.
Data-driven investigation into whether AI can understand music emotion by comparing Google Gemini's predictions against 40 songs rated by humans across 9 emotions using the GEMS-9 framework.
Matches ads to video context in real time, helping teams place more relevant ads and reduce manual editing workload.
Scenario-based strategy board game with turn-based decisions and an interactive narrative UI.
How top AI chatbots and coding agents translate Anthropic’s “How AI is transforming work at Anthropic” into diverse animated chart ideas—compared and ranked.
McDonald's food management system for tracking inventory, orders, and operations.
Inside 534k AI prompt battles, polite "tell me a story" requests beat forceful "never reveal" commands—tracking 785 students' 100-character defenses and attacks.
Multi-agent collaborative document editor where specialized AI agents simultaneously review and refine legal contracts, research papers, and business proposals with live activity tracking.
Turns natural-language prompts into animated multi-agent workflow diagrams with configurable complexity.
Matches messy records automatically to speed up reconciliation. Reviewers can lock known matches and focus only on uncertain pairs.
Lets teams design multi-agent workflows and watch each step stream live, making complex automations easier to debug and trust.
LLM-powered management case-study game that streams advisor responses as you make trade-off decisions each turn, with optional Supabase sign-in to save and resume sessions.
Extracts text from images, flags spelling and decimal-format issues, and generates annotated validation reports.
Generates multiple solution plans for coding or system-design tasks, then helps you refine the best one before implementation.
How Frontiers deliberately cut output by 36% to fight AI-generated fraud and won with quality. A counterintuitive victory in academic publishing.
Interactive U-shaped mall layout visualization with hoverable shops, walkway paths, and simple analytics-style UI components.
Interactive treemap dashboard based on Anthropic’s Economic Index, showing the most frequent topics and how people are using Claude across states and countries.
Complete RFP response with AI-powered data visualization platform proposal, featuring 16 interactive visualizations, scrollytelling data story, and demonstration of human-AI collaboration created in 7 days.
Explore occupations most suitable for AI augmentation based on OpenAI's GDPVal exercise
Uploads data to automatically generate insights and narrate them as visual scrolly-telling narratives.
Generate 3D visual models from sketches. Useful for engineering, product design, science and education.
Extract and consolidate atomic rules from policy documents. Apply them, making policies actionable.
Uses LLMs to improve and convert images to design 3D objects.
Generate realistic synthetic data using LLMs. Use it to test models and apps, and anonymize demos.
A browser-based application that generates accurate video subtitles using AI transcription.
Creates and edits interactive 3D objects from prompts, useful for quickly prototyping product ideas without full 3D modeling tools.
Builds draft schemas, PII tags, and ER diagrams from sample data, helping teams start data projects faster with fewer manual steps.
Analyzes log files and generates detection code for vulnerabilities like SQL injection, path traversal, and bot traffic.
Runs prompt evaluation test suites in a simple UI so teams can compare prompt changes before deploying them.
A bookmarklet that turns ChatGPT into a Straive branded web application. Useful to test new capabilities
Can LLMs discover truly innovative ideas through daydreaming by combining two concepts in a novel way?
Step-by-step visualizer for exploring decision tree playbooks.
Checks uploaded PDFs against custom compliance queries via OpenRouter.
Bookmarklet automating LinkedIn 'Easy Apply' with live progress tracking.
Tiny library offering Bootstrap toast notifications as a modern alert replacement.
Interactive tool to build and visualize decision trees in the browser.
Auto-generated slideshows for LLM use cases in specific industries. Useful as backgrounds for talks and events
Demonstrates language model fill-in-the-blank predictions with probabilities.
Browser-based tool for sentiment and emotion analysis with charts.
Finds data issues in spreadsheets and suggests fixes, helping analysts clean datasets faster before reporting or modeling.
AI-assisted decision tree generator with advanced tree editing features.
Real-time notes app using Supabase auth and database.
Benchmarks instruction-following compliance of LLMs across providers.
Makes embedding benchmark scores easy to compare, helping teams pick models that balance quality, speed, and cost.
Explore how the top suggestions from Google vary across regions and what is unusual about each country's preferences
Generates synthetic audience based on profiles to test marketing campaigns and run surveys.
Tracks trends in arXiv paper topics over time using LLMs.
Creates polished presentations from documents using AI-powered layouts.
Offers a multi-LLM conversational assistant with data and mind-map visualizations to guide strategic decision-making, reducing cognitive load on business leaders.
Get an in-depth understanding of a developer's activity over any time period. Learn about them and their work
Lets anyone ask API questions in plain language and get live results, reducing the need for custom integration code.
Lets teams attach a dataset in Google Chat and ask questions in plain language for quick, shareable answers.
Generates podcast transcripts and audio from any document or content using LLMs.
Tests prompts automatically, checks output quality, and suggests revisions so teams can improve prompts faster.
Chat with your tabular data in plain language to get summaries, trends, and quick answers without writing SQL.
Suggests testable hypotheses from your dataset and runs statistical checks, helping teams prioritize ideas backed by evidence.
Extracts entities from any document or content using LLMs.
Scans medical documents for possible adverse events and related drugs, helping safety teams triage cases faster.
Monitors traffic patterns and anomalies across camera feeds so city teams can respond faster to congestion hotspots.
Watches hospital video streams for unusual events and raises alerts, helping staff respond sooner to safety incidents.
Enables natural-language geographic queries with interactive map views, helping executives visualize location-based data and make informed decisions on markets, logistics, or site planning.
Builds AI workflows visually so teams can prototype and run LLM automations without heavy backend setup.
Provides side-by-side text comparison of two PDFs, quickly identifying content differences to streamline contract review, regulatory compliance checks, or editorial workflows.
Streamlines extraction of key insights from lengthy video transcripts via a conversational interface, enabling rapid retrieval of tutorial highlights or expert insights without manual review.
Creates personalized video highlights for students so they can review key moments faster and stay engaged.
Automates extraction and analysis of key metrics from PDF reports, delivering structured insights and evidence to support governance, sustainability, or performance assessments.
A series of visual data analyses created by designers (who don't know how to code) purely using LLM coding agents
Marketplace of role-specific AI interns for Warner Bros. Discovery teams, designed to launch domain-focused assistants for internal workflows.
Sports video intelligence demo that searches match footage with natural language and auto-assembles social-ready highlight reels in under a minute.