Zum Inhalt springen
NEUFehler-Doktor und Wachturm: fehlgeschlagene n8n-Läufe in einfachen Worten erklärt und Ihnen gemailt. Neuigkeiten ansehen

Bibliothek / Dokumente, PDF & OCR

Dokumente, PDF & OCR

Deep Research Report Generation Using Open Router, Google Search, Webhook/Telegram and

When something happens in a connected app, it reads and extracts data from documents with Claude and Google Gemini using Notion, Telegram.

Was sie bewirkt

# Workflow Overview This workflow automates the process of research planning, query generation, web search, content extraction, and report creation. It integrates with Webhook, Notion, OpenAI, and external APIs. ## Main Steps 1. Trigger: Starts from a Telegram message or Webhook. 2. Strategy Agent: Asks the user for a research topic and clarifies requirements. 3. Query Generation: Generates SERP queries for the research topic. 4. Web Search: Uses Tavily API to fetch search results for each query. 5. Content Extraction: Extracts content from the most relevant URLs. 6. Report Generation: Uses AI to synthesize a detailed report in markdown. 7. Notion Integration: Saves the report and metadata to a Notion database. 8. Notification: Notifies the user when the report is ready. ## Key Nodes - Trigger: Listens for new messages. - Strategy Agent: Handles user interaction and planning. - Search Query Agent: Generates search queries. - HTTP Request: Fetches search results and extracts content. - OpenAI/AI Agent: Processes and synthesizes information. - Notion: Stores the final report. - Sticky Note: Use this to add or update instructions. ## Tips - Update credentials for all API nodes before running. - You can edit or move Sticky Notes for better documentation. - Use the Switch node to handle user confirmations and feedback. - Optionally use Telegram or any other third party integration to trigger the workflow.

Die Schritte

  1. Telegram Trigger
  2. Simple Memory
  3. Structured Output Parser
  4. Switch
  5. Strategy Agent
  6. Search Query Agent
  7. Structured Output Parser1
  8. Split Out
  9. Loop Over Queries
  10. HTTP Request
  11. Edit Fields
  12. OpenAI
  13. HTTP Request1
  14. Aggregate

Insgesamt 37 Schritte.

Ähnliche Automatisierungen