OpenAI + Telegram Morning Digest n8n Workflow Automation

Get a person-checked “morning world digest” in Telegram—powered by OpenAI and Thea

This n8n workflow automation creates a scheduled Telegram digest from Thea ([thealibrary.org](http://thealibrary.org)) and also lets you chat with the shared library via an OpenAI agent—retrieving answers strictly from Thea’s returned entries, with source lines.

What this workflow does

The workflow template provides three independent flows on one canvas:

• 1) Chat with the library
  Uses a Chat Trigger and an AI Agent with three plain HTTP tools (no key): thea_search, thea_find, and thea_entry. You can ask questions like “Who is Karina from aespa?”, “Hotels in 서귀포”, “Any news from Japan since last week?”, or “Is this restaurant in Kuala Lumpur still open?” The agent answers only from Thea results and includes a source line.
• 2) Morning world digest
  A Schedule (08:00) run selects your countries, calls thea_find with since=yesterday for each country, flattens results, makes one LLM call, and sends a compiled digest to Telegram. Thea dates records by the event, so the digest reflects what happened yesterday—deduplicated by people. If a country has no new records, the output includes an honest line.
• 3) Explore a name
  An n8n Form collects a name, runs thea_search, then uses thea_entry to fetch the person’s connected entries (by ids). It generates one entry per neighbor and prompts the LLM with those newest records.

Use cases

• SaaS operators: send daily, source-backed updates to a product team via Telegram.
• Automation engineers: a clean OpenAI + HTTP-tool pattern for grounded Q&A.
• Travel/community teams: compile “what changed yesterday” by country without scraping or RSS.

Technical details

• Nodes / tools: set, code, form, telegram, aggregate, sticky note
• Integrations: OpenAI (AI Agent) + Thea HTTP tools (thea_search, thea_find, thea_entry) + Telegram
• Key logic: grounded responses strictly from Thea returns; schedule at 08:00; date filtering via since.