Pedro Santos
Workflows by Pedro Santos
AI agent web search using SearchAPI & LLM
## 🤖 AI Agent Web Search using SearchApi & LLM ### Who is this for? This workflow is ideal for anyone conducting **online research**, including **students**, **researchers**, **content creators**, and professionals looking for accurate, up-to-date, and verifiable information. It also serves as an excellent foundation for building more sophisticated AI-driven applications. ### What problem does this workflow solve? / Use case This workflow automates web searches by enabling an AI agent to efficiently retrieve and summarize external, verifiable information, ensuring accuracy through source citations. ### What this workflow does * Connects an AI agent node to [SearchApi.io](https://www.searchapi.io/) as an integrated search tool. * Empowers the AI agent to perform real-time web searches using various SearchApi engines (e.g., Google, Bing). * Allows the AI agent to dynamically determine search parameters based on user interaction, delivering contextually relevant results. * Ensures responses include clearly cited sources for validation and further exploration. ### Setup 1. **Install the [SearchApi community node](https://www.npmjs.com/package/@searchapi/n8n-nodes-searchapi)**: * Open **Settings → Community Nodes** inside your self‑hosted n8n instance. * Fill **npm Package Name** with `@searchapi/n8n-nodes-searchapi`. * Accept the risk prompt, and hit **Install**. * It should now appear as a node when you search for it. 2. **API Configuration**: * Set up your [SearchApi.io](https://www.searchapi.io/) credentials in n8n. * Add your preferred LLM provider credentials (e.g., OpenRouter API). 3. **Input Requirements**: * Provide the YouTube video ID (e.g., `wBuULAoJxok`). 4. **Connect LLM Integration**: * Configure the summarization chain with your chosen model and parameters for text splitting. ### How to customize this workflow to your needs * Integrate additional nodes to structure or store search results (e.g., saving to databases, Notion, Google Sheets). * Extend chatbot capabilities to integrate with messaging platforms (Slack, Discord) or email notifications. * Adjust search parameters and filters within the AI agent node to tailor information retrieval. ### Example Usage * **Input**: User asks, "What are the latest developments in AI regulation?" * **Output**: AI retrieves, summarizes, and cites recent, authoritative articles and news sources from the web.
Generate YouTube video summaries with SearchAPI transcripts and LLM
## 🎥 Summarize YouTube Videos using SearchApi & LLM ### Who is this for? This workflow is ideal for **content creators**, **students**, **digital marketers**, **educators**, and **researchers** who want to quickly summarize YouTube videos. ### What problem does this workflow solve? Manually extracting important information from lengthy YouTube videos can be tedious and prone to errors. This workflow streamlines the process by automatically fetching video transcripts using [SearchApi.io](https://www.searchapi.io/) and producing concise, informative summaries through a summarization chain powered by any LLM provider. This allows users to quickly access crucial information without the need for manual transcription or detailed viewing. ### What this workflow does * Fetches the complete transcript of a YouTube video using SearchApi. * Combines the retrieved transcript into a single, continuous text. * Utilizes a **Summarization Chain** with an LLM (e.g., OpenRouter models) to create a concise summary of the video content. ### Setup 1. **Install the [SearchApi community node](https://www.npmjs.com/package/@searchapi/n8n-nodes-searchapi)**: * Open **Settings → Community Nodes** inside your self‑hosted n8n instance. * Fill **npm Package Name** with `@searchapi/n8n-nodes-searchapi`. * Accept the risk prompt, and hit **Install**. * It should now appear as a node when you search for it. 2. **API Configuration**: * Set up your [SearchApi.io](https://www.searchapi.io/) credentials in n8n. * Add your preferred LLM provider credentials (e.g., OpenRouter API). 3. **Input Requirements**: * Provide the YouTube video ID (e.g., `wBuULAoJxok`). 4. **Connect LLM Integration**: * Configure the summarization chain with your chosen model and parameters for text splitting. ### How to customize this workflow to meet your needs * Adjust the summarization model or modify text-splitter parameters to accommodate different lengths and complexities of video transcripts. * Integrate additional nodes to export summaries directly into your preferred tools, such as Google Drive, Slack, or email. * Customize prompt templates in the summarization chain to obtain various summary styles (bullet points, paragraphs, etc.). * Modify the trigger to suit your workflow. ### Example Usage 1. **Input**: YouTube video ID (`wBuULAoJxok`). 2. **Output**: A concise, actionable summary that highlights key ideas, recommendations, and insights from the video.