Museum of Almost: An Interactive Fiction Exhibit Built on Branching Choices

·19 views

The Museum of Almost is a small interactive web exhibition featuring six fictional inventions rooted in everyday objects, such as an umbrella that replays rainy walks and a clock that skips a minute. Visitors navigate the exhibit by making choices — for example, deciding whether to share or keep 'your rain' — which leads them down different narrative branches. The project was built for the Sanity Challenge and launched publicly on September 26, requiring no login to explore. Each invention is clearly labeled as fiction, and the branching stories are pre-authored rather than AI-generated at runtime. A local curator sandbox is also available for developers who want to test the content editing and publication workflow without a Sanity account.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

RELIC Brings Abandoned Codebases Back to Life With Digital Archaeology

RELIC is a new open-source platform that investigates forgotten software repositories to reveal their current condition and how they might be revived. Developers upload a repository as a ZIP archive, and the tool analyzes languages, structure, dependencies, Git history, and unfinished-work markers such as TODO, FIXME, and HACK. Findings are tied to file paths and line numbers, and a locally run Gemma 3 model via Ollama answers natural-language questions grounded in that evidence. The platform then organizes its results into a four-phase revival roadmap and can export a Markdown forensic dossier. Its builder, Ankit, released the project under the MIT License on GitHub as an entry in the Hacktoberfest Open-Source AI Challenge.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How to Automate Pinterest Posting Hourly Without Burning Out

Pinterest favors steady, consistent publishing, and accounts that post in bursts tend to stall once owners get busy with daily business tasks. Posting a fresh pin every hour would mean roughly 720 pins a month, which is hard to sustain manually. The guide outlines a Pinterest automation pipeline that keeps publishing running in the background, along with guidance on safe posting volumes and when to hand the work to a management service. It notes that hourly posting only works when each pin is a genuinely new image with varied titles and descriptions, since Pinterest's spam filters target repetition rather than frequency. It also suggests a catalog of several image variations per product and a range of scheduling tools, from Pinterest's native scheduler to custom API scripts.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Bama Bound AI Video Pricing: What One Clip Costs at 480p vs 720p

A comparison of Bama Bound AI video tools, checked on October 30, 2026, finds a single clip typically costs about $8 to $10 when bought one at a time. The author, who runs bamabound.app, lists $7.99 for 480p and $9.99 for 720p there, paid once with no account required. Other sites such as bamaboundai.com and mimiczip.com sell credit packs or bundles with different clip lengths, which makes direct comparison harder. Subscription-based tools can lower the per-clip cost, but only with a commitment to a plan. Because prices change quickly during trends, readers are advised to recheck pricing before buying.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Tracking flaky Playwright tests across runs to tell flakiness from real breakage

Teams often re-run failing Playwright tests in CI until they pass, which lets unreliable tests pile up and erodes trust in red builds. Playwright's built-in flaky status with retries only covers a single run, so retries can hide problems and a re-run produces a new report that discards earlier failures. Reliable detection requires examining the same test across many runs, looking for passing and failing on the same commit, needing retries to pass, or repeatedly flipping between pass and fail. A test that keeps failing is a broken test needing an immediate fix, which differs from a flaky one that should be stabilized. Identifying a test across runs depends on its spec file, full title path, and project name, so renames or moves can reset its history.

0 comments Read more at DEV Community