Step 2: Define Topics and Locale Inputs

These fields feed the News API query and all AI prompts downstream.

1. Open Define Topics & Locale and add the assignment topics as a string with value AI,n8n.
2. Add the assignment language as a string with value English.
3. Keep Include Other Fields enabled (true) to preserve incoming data for expressions.

Step 3: Connect the News API Request

This node fetches recent articles and passes them into the AI selection step.

1. Open Request News API and set URL to https://newsapi.org/v2/everything.
2. Enable Send Query and add query parameters:
3. Set from to ={{ DateTime.fromISO($json.timestamp).minus({ days: 7 }) }}.
4. Set q to ={{ $json.topics.replaceAll(",", " OR ") }}.
5. Set sortBy to relevancy.
6. Credential Required: Connect your httpQueryAuth credentials in Request News API.
7. Connect Request News API to AI Article Chooser.

Step 4: Select and Split Articles with AI

The workflow uses AI to choose the top five articles and then splits them into individual items.

1. Open AI Article Chooser and confirm jsonOutput is enabled (true).
2. Verify the prompt uses the topic and language expressions: {{ $('Define Topics & Locale').item.json.topics }} and {{ $('Define Topics & Locale').item.json.language }}.
3. Credential Required: Connect your openAiApi credentials in AI Article Chooser.
4. Open Separate Items and set Field To Split Out to message.content.articles.
5. Ensure AI Article Chooser → Separate Items is connected as shown in the execution flow.

Step 5: Generate Refined Articles with the AI Agent

Each article is expanded using a research agent with a tool and structured output parser.

1. Open Newsletter Writing Agent and set Prompt Type to define.
2. Paste the full instruction text into Text, keeping the expressions for the article fields and language, such as {{ $json["title"] }} and {{ $('Define Topics & Locale').item.json.language }}.
3. Confirm Has Output Parser is enabled (true).
4. Ensure LLM Chat Model is connected as the language model for Newsletter Writing Agent. Credential Required: Connect your openAiApi credentials in LLM Chat Model (credentials are set on the parent node, not the agent tool).
5. Ensure Tavily Search Tool is connected as an AI tool to Newsletter Writing Agent. Credential Required: Connect your tavilyApi credentials in Tavily Search Tool (credentials are set on the parent node, not the agent tool).
6. Verify Structured Output Parser is connected as the output parser with the manual schema shown in the node; no credentials required.
7. Connect Separate Items → Newsletter Writing Agent.

Step 6: Aggregate and Send the Telegram Digest

All refined articles are collected into a single digest and formatted for Telegram.

1. Open Aggregate Results and ensure Fields To Aggregate includes output.
2. Open Dispatch Telegram Update and set Chat ID to [YOUR_ID].
3. Set Text to the full expression provided: ={{ $json.output.map(article => { const title = JSON.stringify(article.title).slice(1, -1); const content = JSON.stringify(article.content).slice(1, -1); const source = JSON.stringify(article.source).slice(1, -1); const url = JSON.stringify(article.url).slice(1, -1); return `*${title}*\n${content}\nSource: [${source}](${url})`; }).join('\n\n') }}.
4. Credential Required: Connect your telegramApi credentials in Dispatch Telegram Update.
5. Ensure the execution flow matches Newsletter Writing Agent → Aggregate Results → Dispatch Telegram Update.

Step 7: Test and Activate Your Workflow

Run a manual test to verify the full chain from trigger to Telegram delivery.

1. Click Execute Workflow to run a manual test from Scheduled Automation Trigger.
2. Confirm Request News API returns articles and AI Article Chooser outputs an articles array.
3. Verify Newsletter Writing Agent outputs structured JSON for each article and Aggregate Results consolidates them.
4. Check Telegram for a formatted digest message from Dispatch Telegram Update.
5. When satisfied, toggle the workflow to Active to enable weekly automation.