Step 2: Connect Google Sheets

Log form inputs and AI outputs into Google Sheets for tracking and review.

1. Open Append Form Row and select the target spreadsheet and sheet (document ID [YOUR_ID], sheet Post via form).
2. Map columns in Append Form Row: Date → {{ $now.toFormat("dd/MM/yyyy") }}, Time → {{ $now.toFormat("HH") }}, Content → {{ $json["Input Information here"] }}.
3. Credential Required: Connect your googleSheetsOAuth2Api credentials in Append Form Row.
4. Open Log Content Sheet and select the document [YOUR_ID] and sheet Tự động.
5. Map columns in Log Content Sheet: Ngày → {{ new Date().toLocaleDateString("vi-VN") }}, Nội dung ngắn → {{ $('Parse LLM JSON').item.json.content }}, Nội dung đầy đủ → {{ $json.text?.trim() ? $json.text : $json.message?.content }}.
6. Credential Required: Connect your googleSheetsOAuth2Api credentials in Log Content Sheet.

Step 3: Set Up AI Processing

Configure the LLM pipeline that summarizes, parses, and expands content into a publishable article.

1. In LLM Summary Builder, keep the Text prompt as provided and ensure it includes the dynamic content block expression (e.g., {{ $json.message?.text?.trim() ? $json.message.text : $json.content?.trim() ? $json.content : $json['Content']?.trim() ? $json['Content'] : "" }}).
2. Ensure Gemini Chat Model A is connected as the language model for LLM Summary Builder.
3. Credential Required: Connect your googlePalmApi credentials in Gemini Chat Model A (credentials should be added to the model node, not the chain node).
4. Open Parse LLM JSON and keep the parsing code that extracts JSON from ```json ... ``` blocks.
5. Parse LLM JSON outputs to both Image Prompt Check and LLM Article Writer in parallel.
6. In LLM Article Writer, keep the Text prompt and ensure it references {{ $json.content }} for the summary input.
7. Ensure Gemini Chat Model B is connected as the language model for LLM Article Writer.
8. Credential Required: Connect your googlePalmApi credentials in Gemini Chat Model B (credentials should be added to the model node, not the chain node).

Step 4: Configure Image Generation and Content Assembly

This step generates optional images, cleans content, and merges text and media for publishing.

1. In Image Prompt Check, confirm the boolean condition uses {{ !!$json.prompt_image }} to trigger image generation only when a prompt exists.
2. Open Image Generation Call and set URL to https://router.huggingface.co/together/v1/images/generations, Method to POST, and JSON Body to { "response_format": "b64_json", "prompt": "{{ $json.prompt_image }}", "model": "black-forest-labs/FLUX.1-schnell" }.
3. Set Header Parameters in Image Generation Call to Authorization: Bearer [CONFIGURE_YOUR_TOKEN] and Content-Type: application/json.
4. In Build Image File, keep the code that converts $json.data[0].b64_json to binary with filename image.png.
5. Open Upload Facebook Image and set Edge to photos, Node to [YOUR_ID], Send Binary Data to true, and Binary Property Name to data. Ensure published is false.
6. Credential Required: Connect your facebookGraphApi credentials in Upload Facebook Image (this node needs credentials but none are configured).
7. Open Clean Content Text and keep the function that removes # and * characters and outputs content and platform.
8. Set Combine Streams to Mode combine and Combine By combineByPosition to merge image and text data.

Step 5: Configure Publishing and Notifications

Publish the post to Facebook and send a confirmation email when the post is successful.

1. Open Publish Facebook Post and set Edge to feed, Node to [YOUR_ID], and Graph API Version to v22.0.
2. In Publish Facebook Post query parameters, set message to {{ $('Clean Content Text').item.json.content }} and attached_media to {{ $json.id ? `[{\"media_fbid\":\"${$json.id}\"}]` : "" }}.
3. Credential Required: Connect your facebookGraphApi credentials in Publish Facebook Post.
4. In Post Result Check, verify the condition uses {{ $json.post_supports_client_mutation_id }} to gate the notification step.
5. Open Send Email Notice and set Send To to [YOUR_EMAIL], Subject to ✅ The draft version is uploaded, and Message to ✅ The post will be published after 2 minutes! 📄 Review in: https://facebook.com/{{ $json.id.split('_')[0] }}/posts/{{ $json.id.split('_')[1] }}.
6. Credential Required: Connect your gmailOAuth2 credentials in Send Email Notice.

Step 6: Test and Activate Your Workflow

Run a full test to confirm AI output, image generation, publishing, and email notifications.

1. Click Execute Workflow and submit the Form Input Trigger with a 50+ word idea.
2. Confirm Append Form Row and Log Content Sheet add rows to their respective sheets.
3. Verify that Parse LLM JSON outputs valid prompt_image and content, and that image generation only runs when a prompt exists.
4. Check Facebook to ensure Publish Facebook Post creates a post, and verify the email from Send Email Notice includes the correct post URL.
5. When everything works as expected, toggle the workflow to Active for production use.