Sort saved bookmarks and review uncertain classifications
At a glance
A documented n8n workflow files Raindrop.io bookmarks with an LLM and sends uncertain items to a review collection. This workflow is listed under Bookmarks, Personal Organization, Human Review.
- Agents / tools
- n8n, Raindrop.io, Ollama, Telegram
SourceBuilding a tool for these workflows? See weekly rates.
What they did
1) Import the template into n8n. Create a Data Table named collections with collection_id, name and description columns and one row per Raindrop destination. 2) Give the Raindrop HTTP nodes a bearer-token credential that can read and update your bookmarks. Configure a reachable Ollama /api/chat endpoint or an OpenAI-compatible /chat/completions endpoint, including authentication if needed. 3) Run the setup form and enter the Inbox and fallback collection IDs, confidence threshold and optional filing rules. Add a Telegram bot credential and alert chat ID, or disable notifications. 4) Run a check, then use the fifteen-minute schedule: it fetches the latest 50 Inbox bookmarks and sends each title, URL and excerpt for a collection ID and confidence score. 5) Review the moved bookmarks. Invalid or low-confidence responses go to the fallback collection and produce a Telegram alert when enabled.