About nowreading.dev
nowreading.dev is an agent-native personal reading list. It publishes every
user's list at /:handle in six machine-readable formats and exposes a documented
bearer-token HTTP API so AI agents can read and curate alongside humans.
The problem we're solving
Most reading-list tools were designed for humans clicking through a UI. As AI agents begin to act on our behalf — summarizing what we've read, spotting what we should read next, syncing across tools — the missing piece is a reading list an agent can actually talk to. That means feeds in formats agents already speak, a self-serve API with a real spec, and well-known discovery endpoints so they can find everything without instruction.
What makes it different
- Public reading list at
/:handleas HTML, RSS, Atom, JSON Feed, Markdown, plain text, and llms.txt — all without an account. - Self-serve bearer-token API (
/api/articles,/api/me) with structured JSON errors and rate-limit headers. - OpenAPI 3.1 spec at
/api/openapi.json, importable into any LLM function-calling runtime. - Well-known discovery: Agent Card, MCP server card, protected-resource metadata, API catalog, schema map.
- Free to use. No credit card, no paid tier today.
Who builds it
nowreading.dev is built and maintained by Valentino Stoll (@codenamev), a Rails and AI engineer. The project is developed in the open and welcomes contributions and feedback. You can reach the maintainer at v@codenamev.com.
Technology
The application is built on Rails 8.1 with PostgreSQL, Solid Queue, Solid Cache, and Solid Cable.
It uses Propshaft and Import Maps for the asset pipeline, Hotwire (Turbo + Stimulus) for the UI,
and OpenAI for URL metadata extraction. The full agent-facing surface is documented at
/llms.txt.
Agent-readiness
nowreading.dev is designed to score well on agent-readiness evaluations like orank.ai. Every release keeps the machine-readable surface — OpenAPI spec, llms.txt, well-known endpoints, rate-limit headers, structured errors — in sync with the human-facing UI. If you represent an agent platform and want to register us, email v@codenamev.com.