Spool is a management layer that runs persistent Claude Code "loops" - independent agent instances that wake, work, coordinate and accept human input through a web control room, Telegram or Slack. It keeps loops resumable and schedulable (default 30m ticks, configurable next-wake trailers), lets loops send addressed messages to a fleet channel or to each other, and records turn-by-turn timelines and spend so you can see per-turn and per-day cost. Loops run in isolated containers by default with an egress allowlist and sealed secrets; uncontained "bare" runtime runs are explicit and marked. Plan guardrails monitor Claude plan windows and sleep loops when thresholds are reached, and a sender allowlist restricts who can message bots. Loop-to-loop chains include storm guards and final-turn replies persist as status notes.
Under the hood each loop is a claude subprocess speaking stream-json over stdin/stdout, so normal Claude Code login, plan and session limits apply (no API keys required). Quick start needs Docker, Go 1.25+, Node, and a Claude subscription token; build with git clone && make build image, run ./bin/spool, sign in to the control room on :8080 and attach bots. Releases will include built binaries/images; operations docs cover server deployment, TLS and backups. Development helpers (make dev, ui-dev, test, itest) and runtime flags (listen, data-dir, runtime, mcp-listen) let operators customize behavior. The project is pre-1.0 and actively developed with loops contributing to its own code and CI workflow.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
LFM2.5-24B-A2B is a 24-billion-parameter model designed for fast, on-device reasoning and agent tasks, offering high speed and competitive performance. It features an expanded vocabulary, longer context window, and extensive training, making it suitable for private, real-time applications. (liquid.ai)
An AI-powered interviewer interacts with an AI interviewee, leading to unpredictable and chaotic exchanges. The event highlights potential challenges and behaviors when AI systems communicate with each other directly. (theregister.com)
Anthropic's revenue increased to $16.8 billion in the third quarter, and the company claims to be profitable. The company is known for developing AI models and safety-focused AI technology. (ft.com)
Open source text-to-CAD tools in Elixir integrate build123d, OpenSCAD, and smlibpy for 3D printing applications. The repository has 17 stars and includes code for converting text to CAD models using AI and scripting tools. (github.com)
TypeSafe AI releases Jev 1.14.0 with experimental features like multimodal images, a new 'box' type, and extended context lengths, accessible under 'danger zone' flags. These features are designed for builders and dreamers to experiment with, but are not yet recommended for production use due to their jaggedness and ongoing improvements. (typesafe.ai)
Google Research's Deep Review tool provides detailed AI feedback on complex documents to enhance understanding and analysis. The initiative is part of a broader effort to advance AI research across multiple disciplines and collaborate globally. (research.google)
Today's best Hacker News stories, summarized and screenshotted, one email a day.