Raindrop launches Stella: a private AI assistant for bookmark discovery and organization

Raindrop has released Stella, an artificial intelligence assistant now in beta for its web app. Stella enables users to engage with their bookmark collections, saved articles, and documents using conversational language. Functioning as a private large language model, Stella analyzes only material users have saved, ensuring responses are grounded in the user's own content without input from external sources.

Following the introduction of Stella, users no longer need to recall exact titles or URLs. Instead, they can describe what they remember and Stella will locate the relevant content. The assistant also summarizes lengthy articles, explains difficult topics using plain language, allows comparison of two pieces, and previews the value of articles before a full read.

Adding to its capabilities, Stella assists with organizing bookmarks. Users can instruct it to sort unsorted folders, clear duplicate tags, move items to the proper collections, or suggest structural improvements for their library. Stella proposes changes, but users retain control with final approval.

From a privacy standpoint, Stella runs exclusively on Raindrop’s own servers using an open-source GPT OSS model. All communication is encrypted with TLS 1.2 or higher, and data never leaves Raindrop's infrastructure or goes to outside AI vendors. AI processing is only initiated by users, with no background analysis or automatic opt-in.

Stella is currently available to Raindrop Pro subscribers through the web app.