MAKE YOURSELF AT HOME

Settings_

Your models, sources of discovery, and reading preferences.

Models

Classification uses Jev. Summaries and article conversations use your OpenCode Go subscription.

Summary modelDefault chat model

API connections

Keys stay on this machine. Leave a field blank to keep its existing key.

TypeSafe AI / Jev · connected via .envOpenCode Go · connected via .env

Discovery

Followed blogs, Hacker News, and links in strong articles bring new reading into your library. Run the local worker for scheduled discovery.

Discover automaticallyInclude Hacker NewsHours between discovery runsArticles processed per worker pass

Reading preferences

Likes and dislikes gradually shape your recommendations. You can also set explicit weights from −3 to 3. Use topic names, technology names, source domains, or article types. Exploration stays between 0.2 and 0.6.

Preference profile (JSON)

Library operations

Discovery queues candidates. Processing extracts and scores them. Only useful articles enter your recommendations.

Recent activity

9/19/2026, 11:56:52 PM UTC　ingest / complete

{"processed":1,"recommendations":12}

9/19/2026, 11:53:51 PM UTC　ingest / complete

{"processed":1,"recommendations":12}

9/19/2026, 11:51:03 PM UTC　ingest / complete

{"processed":8,"recommendations":12}

9/19/2026, 11:49:29 PM UTC　ingest / complete

{"processed":8,"recommendations":12}

9/19/2026, 11:47:39 PM UTC　ingest / complete

{"processed":8,"recommendations":12}

9/19/2026, 11:47:33 PM UTC　discover / complete

{"sources":8,"candidates":30}

9/19/2026, 11:44:49 PM UTC　ingest / complete

{"processed":1,"recommendations":12}

9/19/2026, 11:44:28 PM UTC　ingest / complete

{"processed":5,"recommendations":12}