Sources
The Wider Lens reads only public sources, server-side. Headlines and short excerpts are shown; full stories live with their original publishers. Sources can be added or removed inlib/sources/registry.tswithout touching the rest of the app.
News feeds (RSS/Atom)
- Visit site↗
OpenAI News
official feed
- Visit site↗
Google DeepMind Blog
official feed
- Visit site↗
Google Blog — AI
official feed
- Visit site↗
Hugging Face Blog
official feed
- Visit site↗
NVIDIA Blog
official feed
- Visit site↗
Mistral AI News
official feed
- Visit site↗
Microsoft Research Blog
official feed
- Visit site↗
Stability AI News
official feed
- Visit site↗
Transformer Circuits
official feed
- Visit site↗
TechCrunch — AI
journalism feed
- Visit site↗
The Verge
journalism feed
- Visit site↗
MIT Technology Review — AI
journalism feed
- Visit site↗
Ars Technica
journalism feed
- Visit site↗
Wired — AI
journalism feed
- Visit site↗
MarkTechPost
journalism feed
- Visit site↗
The Decoder
journalism feed
- Visit site↗
Towards Data Science
journalism feed
- Visit site↗
Import AI
journalism feed
APIs
arXiv API research
Categories: cs.AI, cs.LG, cs.CL, cs.CV, cs.RO, stat.ML. Queried at most once per page render; results cached for 60 minutes.
GitHub REST API github
Curated AI topic searches plus latest releases. Works without a token (60 requests/hour); set
GITHUB_TOKENfor 5,000/hour. Cached 30 minutes.Hugging Face Hub API huggingface
Public model listings sorted by recent update and by likes. No key required. Cached 60 minutes.
OpenRouter API models
Public model directory with real release timestamps — powers the “Newly released” model list. No key required. Cached 60 minutes.
What we don't do
- No scraping of sites without public feeds or APIs.
- No invented statistics — star counts, likes, and downloads come straight from the APIs.
- No full-article republication — excerpts only, always linking to the original.
