Step 2: Connect Google Sheets

Pull recent ideas and append new posts to your sheet.

1. Open Retrieve Sheet Rows and select the target spreadsheet and sheet.
2. Credential Required: Connect your googleSheetsOAuth2Api credentials in Retrieve Sheet Rows.
3. Open Append Sheet Record and choose the same spreadsheet/sheet for output.
4. Credential Required: Connect your googleSheetsOAuth2Api credentials in Append Sheet Record.
5. In Append Sheet Record > Columns, keep the mappings as configured, including idea to {{ $('Flatten Output').item.json.idea }} and output to {{ $json.output.post }}.

Step 3: Limit and Research Ideas

Use recent ideas to avoid duplication and generate new research-backed topics.

1. In Restrict Items, set keep to lastItems and maxItems to 5.
2. Open Perplexity Research Query and keep model set to sonar-pro.
3. Verify the system prompt references the last five ideas using expressions like {{ $('Restrict Items').all()[0].json.idea }} to avoid repeats.
4. Credential Required: Connect your perplexityApi credentials in Perplexity Research Query.

Step 4: Set Up AI Interpretation and Post Generation

Parse ideas, write LinkedIn posts, and enforce structured JSON output.

1. Open OpenAI Chat Engine and select model gpt-4.1-mini with the provided options.
2. Credential Required: Connect your openAiApi credentials in OpenAI Chat Engine.
3. In Interpret Ideas, set text to {{ $json.choices[0].message.content }} and keep hasOutputParser enabled.
4. Ensure Idea JSON Parser is attached as the output parser for Interpret Ideas. This parser is a sub-node; credentials are handled by OpenAI Chat Engine, not the parser itself.
5. In Compose LinkedIn Post, set text to Generate a post based on this: Idea: {{ $json.idea }} Details: {{ $json.details }} Style: {{ $json.style }} Your output must always be in the following JSON format: { "post": "" } and keep hasOutputParser enabled.
6. Ensure Post JSON Parser is attached as the output parser for Compose LinkedIn Post. This parser is a sub-node; credentials are handled by OpenAI Chat Engine, not the parser itself.

Step 5: Transform Output and Send Notifications

Flatten the AI output array and alert your team after data is appended.

1. In Flatten Output, keep jsCode set to return $input.first().json.output.flat(1).
2. Confirm the flow Interpret Ideas → Flatten Output → Compose LinkedIn Post → Append Sheet Record.
3. Open Telegram Alert and set chatId to your Telegram ID, replacing [YOUR_ID].
4. Keep the notification text as New ideas for LinkedIn have been generated!.

Step 6: Test and Activate Your Workflow

Validate end-to-end execution and turn on the schedule.

1. Click Execute Workflow to run a manual test from Timed Schedule Start.
2. Confirm Append Sheet Record adds new rows with values mapped from Flatten Output and Compose LinkedIn Post.
3. Check that Telegram Alert sends the notification message after the sheet append completes.
4. Toggle the workflow to Active to run automatically at the scheduled time.