AnyoneMakes

AI-assisted creation examples

See where AI fits into making a visual novel and a protein-research workflow, and what each project delivers.

Using AI to build a project does not mean the finished product runs AI. The visual-novel creator describes help with art and writing; Extremolith has an event description and a public repository. Neither case alone tells us how well the finished work performs.

2 featured examples · 61 related projects

Compare the inputs and outputs

On smaller screens, scroll sideways to see the full comparison.

Based on the sources below, not a ranking of results.
ProjectInput or production taskOutput
When the Keys FallWriting and artwork for a visual novelA Chinese visual novel with a demo link, plus flowcharts and a CG gallery for revisiting the story.
ExtremolithProtein sequences and environmental dataEnvironment-labeled data tables, protein sequence files and a page showing a fixed, precomputed laccase result.

When the Keys Fall

Use AI to help write and illustrate a playable visual novel.

How it works

  1. The creator reports using AI to help with the artwork and script.
  2. Build the story around branching choices, flowcharts and simulated-phone interactions.
  3. Offer a PC demo through the public product page.

What you get

A Chinese visual novel with a demo link, plus flowcharts and a CG gallery for revisiting the story.

AI-assisted production does not establish live AI conversation inside the game. We have not played the full story or tested save behavior.

Sources last checked: 2026-10-10

Extremolith

Use Claude to help build a data and model workflow for protein research.

How it works

  1. Combine GTDB protein and environmental data with GenomeSPOT predictions.
  2. Filter secreted proteins and split the training data.
  3. Process the data using the repository commands; the event description also covers training and generation.

What you get

Environment-labeled data tables, protein sequence files and a page showing a fixed, precomputed laccase result.

The page does not perform live inference from user input. We did not run the code, train a model or carry out biological experiments, so this is not a performance test.

Sources last checked: 2026-10-21

Explore how projects are built

See what goes in, how it works and what comes out.

Browse all projects with this tag

Projects with this tag may serve different needs. A shared tag does not mean they have the same features or deliver the same results.

61 projects · Page 1 of 3

Kulimi

A light action puzzle in which clicking a ball causes an explosion whose blast pushes another ball into a hole. The creator describes about 20 levels, AI-generated background music, and a 5–15-minute playthrough; the page also lists half an hour, so duration is inconsistent and unmeasured.

OINK

An AI-animated narrative short film, recorded in the event as a Best Narrative winner. Its theme concerns the tension between illusion and reality, and whether happiness built on an illusion can survive reality’s intrusion.

Zeppelin

Zeppelin is an S3-native vector search engine that separates storage from computation. The creator describes storing vectors in object storage and using stateless, coordination-free query nodes to scale horizontally, rather than maintaining replicated in-memory index clusters.

The Wonderful Expedition

A strategy roguelike combining auto-battling and resource management. A commander recruits units and builds synergies by faction, class or unit combinations while clearing monsters along an expedition. The creator lists four factions, seven classes, 28 unit types and 11 treasures, with AI-assisted art and music; these counts were not independently verified.