Arquivo Fantasma Turns Forgotten Digital Fragments into Micro-Horror Stories

·1 views

Arquivo Fantasma is a local-first command-line tool that generates short eerie stories from user-supplied text fragments. It offers three narrative styles: horror, folklore, and cyberpunk, with the cyberpunk style focused on haunted repositories and abandoned systems. The project is written in Python, has no mandatory third-party dependencies, and can optionally use a locally installed open-weight model through Ollama. Its design prioritizes privacy, since no hosted AI service or API key is needed and only explicitly selected text is processed. A deterministic offline fallback keeps it usable without any model or network connection.

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)

Tailscale Shares Update on SSH Investigation and Remediation Efforts

Tailscale has published an update on its investigation into an issue involving its SSH feature. The post, shared on Hacker News, outlines the company's progress in examining the matter and the steps taken to address it. The article text available is limited to the title and link, so specific details about the nature of the problem and its scope are not included in this summary. Readers seeking the full findings and remediation measures can refer to Tailscale's original post.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

Team Eliminates Support Tickets With AI Bot Trained on Markdown Docs at Zero Cost

A development team describes how they replaced support tickets with an AI bot trained on their Markdown documentation. The article is published on DEV Community and was posted on October 23. The bot was built so that users can get answers from the existing docs without submitting a ticket. The authors state that the setup runs at $0 per month. The source text is brief and provides few further technical details about the implementation.

0 comments Read more at DEV Community

ProgrammingHacker News ·

Survival LLM: Offline AI Assistant and Emergency Library for Your Own Computer

A new project called Survival LLM has been shared on Hacker News, presenting an AI assistant designed to run locally on a user's computer. It is also described as an emergency library, suggesting it bundles reference material for use when internet access is unavailable. The post drew 24 points and 8 comments on the discussion forum. The available source details are limited to the project's web page and the discussion thread, so specific features and technical details are not confirmed here.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

Choosing Retry Boundaries for Node.js Worker Pools: SQS, RabbitMQ, or BullMQ

The article argues that the key question for a rate-limited worker pool is who owns a failed delivery and what happens after the final attempt, not which queue is fastest. Amazon SQS suits small teams that want managed operations and a policy-driven dead-letter queue for repeated failures. RabbitMQ offers broker-level control and routing flexibility, but the team must build its own retry and dead-letter logic. BullMQ provides a Node.js-oriented job API with explicit retry and backoff, though it makes Redis part of the reliability boundary. In every option, the author recommends making the worker idempotent to avoid duplicate updates and endless requeue loops.

0 comments Read more at DEV Community