Get early access

Personal feed & knowledge base

Stop monitoring.
Start knowing.

feedql watches the keywords, researchers, repositories, forums and registries you care about — then removes the duplicates, groups reports of the same event and leads you to the original source.

Like Ahrefs or Sensor Tower — only for news. No spam, one email when your seat is ready.

feedql
topic "open-weight LLMs" from hn, reddit, github, x since 24h min_score 0.7
  1. 0.94
    New 70B model released under Apache-2.0 14 reports folded · first seen on GitHub · 2h ago
  2. 0.88
    Benchmark dispute: eval set leaked into training data 6 reports folded · first seen on Hacker News · 5h ago
  3. 0.81
    EU publishes guidance for general-purpose AI models 3 reports folded · source: official journal · 9h ago

One place instead of a zoo of scripts.

Keeping up in the age of AI means tracking a dozen kinds of signal at once. Done with separate scripts, alerts and cloud automations, it turns into a zoo of tools that each break in their own way. feedql keeps it in one deduplicated pool you can read and query.

How it works

  1. 01

    Collect

    Pick keywords, people, repositories, forums and registries. feedql polls each source on its own schedule and respects its limits.

  2. 02

    Deduplicate

    The same story from twelve places becomes one item. Reposts, mirrors and rewrites fold into the original.

  3. 03

    Cluster & trace

    Reports about one event are grouped, ranked against your topics, and linked back to the first source.

  4. 04

    Read or pipe

    Read it on the web, or pull it through the API straight into your own tools and AI agents.

Sources it already reads

Built for people who have to stay on top of their field

Researchers, founders, analysts, investors, policy and product teams — anyone who needs a steady pulse on a topic without monitoring it by hand.

Get early access

We are opening feedql to a small group first. Leave your email and tell us what you follow.