Create and scale AI content steps with a clear structure. Teams can send one request from an app or a batch of prompts and receive organized outputs for blogs, emails, or product copy. The goal is faster drafting with consistent tone and easy review.

Input arrives by a manual test or a webhook. Prompts are prepared in a Set node and reshaped into a clean list of instructions. You can run four Anthropic steps in sequence with memory for context, or fan out prompts in parallel for speed. A memory node keeps the last ten turns and a clean step resets it when needed. A web page can be fetched and converted to markdown as optional source material. Merge nodes collect every result and align them with the original prompts.

You need an Anthropic API key and a working webhook URL. Choose the path you prefer by connecting the labeled links to the sequential chain or the parallel block. Expect faster drafts and more consistent messages across channels. Strong fits include content calendars, email variants at scale, landing page copy, and outline to article expansion.

• Manual test and webhook intake so you can trigger from an app or run safe dry runs in the editor.
• Set and Reshape nodes build a clean list of step instructions from your fields.
• Sequential chain of four Anthropic Chat Model nodes with temperature set to 0.5.
• Parallel processing path to fan out prompts for speed when memory is not needed.
• Simple Memory buffer keeps the last 10 turns and a Clean memory node can reset the session.
• HTTP Request fetches a page and Markdown converts HTML to markdown for the model.
• Merge nodes assemble all outputs and align them with the original prompts.
• Webhook URL ready for external systems and forms to send content requests.

• Reduce draft creation time from 2 hours to 10 minutes
• Automate up to 80 percent of repetitive prompting tasks
• Handle up to 4 times more requests with parallel runs
• Keep context across 10 turns for more consistent tone
• Connect forms or apps to AI with a simple webhook
• Unify outputs in one combined stream for quick review

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Anthropic and Webhook. See the Tools Required section above for links to create accounts with these services.
3. Open the Anthropic Chat Model nodes and create a new credential: double click the node, in the 'Credential to connect with' dropdown click 'Create new credential', then follow the on screen instructions. In your Anthropic account, create an API key on the API page and paste it into n8n. Name the credential clearly, for example 'Anthropic Prod'.
4. Open the Webhook node and copy the Production URL. If you use n8n Cloud, replace any environment variable in the URL with your actual instance URL as noted in the sticky note. No separate credential is needed for the webhook.
5. Enter your prompts: in the 'Initial prompts' and 'Initial prompts1' Set nodes, fill the system_prompt and the step fields you want the model to follow. Keep names short and clear.
6. Choose processing mode: connect the labeled CONNECT ME nodes to either the sequential LLM chain for memory or the parallel block for speed. Keep only one path active while testing.
7. Memory settings: in the Simple Memory node, keep sessionKey as a fixed value for shared context, or change it if each request should have its own memory. Use the Clean memory node to reset all context when needed.
8. Optional source content: in the HTTP Request node, set the URL to your page. The Markdown node will convert the HTML to markdown and make it available to the model.
9. Test the webhook: send a POST with JSON that includes your fields, for example a system_prompt and several step fields. Confirm that Split Out and Merge nodes process each item.
10. Run a manual test: click Execute Workflow and review the outputs from the sequential or parallel branch. Check that the final merge aligns outputs with your original prompts.
11. Troubleshoot: if no response returns, confirm the webhook URL matches the one you called. If model calls fail, verify your Anthropic API key and project limits. If outputs mix between runs, reset memory with the Clean memory node.