Step 2: Connect SharePoint data sources and files

These nodes load session cookies, profile lists, and previously commented posts, then update the tracking file.

1. Open Load Session Cookies and select the file instagram_session_cookies.txt in your SharePoint site and folder.
2. Credential Required: Connect your microsoftSharePointOAuth2Api credentials in Load Session Cookies.
3. Open Retrieve Account List and confirm File Name is instagram_profiles_to_scrape.csv.
4. Credential Required: Connect your microsoftSharePointOAuth2Api credentials in Retrieve Account List.
5. Open Download Commented File and select instagram_posts_already_commented.csv.
6. Credential Required: Connect your microsoftSharePointOAuth2Api credentials in Download Commented File and Update Commented File.

⚠️ Common Pitfall: Ensure the SharePoint folders and file IDs match across Load Session Cookies, Retrieve Account List, Download Commented File, and Update Commented File to avoid file-not-found errors.

Step 3: Set up session cookie selection and environment variables

This section extracts the cookies and selects one based on Berlin time, then stores values for downstream agents.

1. Open Extract Cookie Text and set Operation to text.
2. Open Choose Session Cookie and verify the prompt includes the expressions {{ $now.setZone('Europe/Berlin').format('DD HH:mm:ss') }}, {{ $now.setZone('Europe/Berlin').format('HH') }}, and {{ $json.data }}.
3. Open Assign ENV Values and set ENV_SESSION_COOKIE to {{ $('Choose Session Cookie').first().json.output.parseJson().session_cookie }}.
4. In Assign ENV Values, set ENV_COMMENT_PROMPT to Erstelle einen ansprechenden Instagram-Kommentar basierend auf dem gegebenen Post-Inhalt. and ENV_COMMENT_LANGUAGE to Deutsch.

Credential Required: Connect your openAiApi credentials in OpenAI Chat Engine B. This model is linked to Choose Session Cookie, so credentials must be added to the parent model node.

Step 4: Configure Phantombuster extractor and post list retrieval

This chain fetches a profile extractor, runs it, waits, and downloads a filtered list of post URLs.

1. Open Fetch Profile Extractor and set Agent ID to your Phantombuster agent ID (replace [YOUR_ID]).
2. Credential Required: Connect your phantombusterApi credentials in Fetch Profile Extractor, Run Extractor Agent, Primary API Request, and Secondary API Request.
3. Open Run Extractor Agent and ensure Agent ID is set to {{ $('Fetch Profile Extractor').item.json.id }}.
4. In Run Extractor Agent, verify Arguments JSON includes "spreadsheetUrl": "{{$('Retrieve Account List').item.json['@content.downloadUrl']}}" and "sessionCookie": "{{$('Assign ENV Values').item.json.ENV_SESSION_COOKIE}}".
5. Open Pause Gate 3 and set Amount to 30 seconds.
6. Open Primary API Request and set URL to https://api.phantombuster.com/api/v2/agents/fetch?id=[YOUR_ID].
7. Open Secondary API Request and set URL to =https://phantombuster.s3.amazonaws.com/{{ $json.data?.orgS3Folder || $json.orgS3Folder }}/{{ $json.data?.s3Folder || $json.s3Folder }}/filtered_result.csv.
8. Open Prepare Post Links and confirm the code parses the CSV into { postUrl } objects.
9. Open Empty List Check and ensure the condition is {{ $json.isEmpty() }} equals true to skip empty lists.

⚠️ Common Pitfall: If the filtered CSV is empty, Empty List Check routes back to Pause Gate 2 and no comment is posted. Verify your extractor agent produces URLs.

Step 5: Prevent duplicates and update the commented posts list

This block avoids commenting on posts that were already processed, and updates the tracking CSV after new comments.

1. Open Download Commented File and ensure it loads instagram_posts_already_commented.csv.
2. Open Parse File Records and keep Header Row enabled.
3. Open Detect Duplicate Post and confirm it compares {{ $('Pick Random Post').first().json.postUrl }} against existing records.
4. Open Duplicate Check Gate and ensure the condition is {{ $('Detect Duplicate Post').first().json.isDuplicate }} equals true.
5. Open Assemble Updated List to append the new post URL to the list.
6. Open Convert Records File then Update Commented File and confirm File Name is instagram_posts_already_commented.csv with Change File Content enabled.

Step 6: Generate the comment and prepare the upload CSV

These nodes generate a short German comment, then turn it into a clean one-row CSV file for the Phantombuster comment agent.

1. Open Pick Random Post and keep the code that selects postUrl and description from the list.
2. Open Generate Comment and keep the prompt text using {{ $('Assign ENV Values').first().json.ENV_COMMENT_PROMPT }}, {{ $('Assign ENV Values').first().json.ENV_COMMENT_LANGUAGE }}, and {{ $('Pick Random Post').first().json.description }}.
3. Credential Required: Connect your openAiApi credentials in OpenAI Chat Engine. This model is linked to Generate Comment, so credentials must be added to the parent model node.
4. Open Build CSV Binary and confirm the output file name is phantombuster_clean.csv.

Step 7: Upload CSV and run the AutoComment agent

This section uploads the CSV to SharePoint, runs the comment agent, and retrieves its output after a short wait.

1. Open Upload CSV File and set File Name to instagram_post_to_comment.csv with Operation upload.
2. Credential Required: Connect your microsoftSharePointOAuth2Api credentials in Upload CSV File.
3. Open Fetch AutoComment Agent and set Agent ID to your Phantombuster agent ID (replace [YOUR_ID]).
4. Open Run AutoComment Agent and confirm Arguments JSON uses {{ $('Assign ENV Values').first().json.ENV_SESSION_COOKIE }} and {{ $('Upload CSV File').item.json['@content.downloadUrl'] }}.
5. Open Pause Gate 1 and keep Amount at 30 seconds before Retrieve Agent Output.
6. Credential Required: Connect your phantombusterApi credentials in Fetch AutoComment Agent, Run AutoComment Agent, and Retrieve Agent Output.

Step 8: Test & Activate Your Workflow

Validate the end-to-end flow before scheduling it for production use.

1. Click Execute Workflow and confirm that Scheduled Run Trigger starts the run and Load Session Cookies loads the cookie file.
2. Check that Generate Comment returns an output string and Build CSV Binary creates phantombuster_clean.csv.
3. Verify Upload CSV File uploads instagram_post_to_comment.csv, and Run AutoComment Agent runs successfully.
4. Confirm Update Commented File updates instagram_posts_already_commented.csv after a successful comment cycle.
5. When satisfied, toggle the workflow to Active to enable the scheduled runs.