Summarize company profiles from domains with ReplyNodes and OpenAI

Quick overview

This workflow researches a list of company domains by fetching website content through ReplyNodes endpoints, then uses an OpenAI-compatible chat model to extract a structured company profile. Each row includes a status and evidence URLs.

How it works

1. Starts when you run the workflow manually.
2. Loads endpoint, rate-limit and text-size settings, then normalizes the domain list and processes it one domain at a time.
3. For each valid domain, fetches the homepage as Markdown via ReplyNodes and extracts key site links (pricing, docs, careers, and social profiles).
4. If the homepage content is usable, fetches the pricing page Markdown, brand JSON data, and checks whether a logo is available via ReplyNodes.
5. Combines the collected website content into a single prompt and sends it to a chat LLM (OpenAI model by default) to generate a structured company profile.
6. Builds a final output row with the LLM fields plus discovered links, logo URL, brand colors, evidence URLs, and a status/error value, then waits before continuing to the next domain to respect public rate limits.

Setup

1. Add credentials to the chat model node. The workflow uses OpenAI by default; you can replace it with any chat model supported by n8n's Basic LLM Chain (only OpenAI-compatible models were tested).
2. Edit the domain list in the Domains code step to include the companies you want to research.
3. (Optional) Adjust the wait time and maximum captured text sizes in the Settings code node. To stop sending the X-Source attribution header, delete it in the four HTTP Request nodes: Fetch homepage, Fetch pricing, Fetch brand, and Check logo.

Requirements

• An LLM credential is required; no data-provider API key is needed. ReplyNodes endpoints work without a key, have public limits of about 20 requests per minute per IP, and are operated by the template author. The workflow does not return headcount, funding, or contact emails.