# Omnivore (Read-Later) MCP for AI Agents AI Agent Connect > Omnivore (Read-Later) MCP lets you manage your saved reading list directly through your AI agent. You can search for specific articles by label or status, pull full content for analysis, and save new links to your library without switching tabs. It turns your bookmarks into an actionable knowledge base. ## Overview - **Category:** productivity - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_e0lCuq3YhluUqdYO6MN37sEg8a31l16jxuwtJbtt/ai-agent-connect - **Tags:** read-it-later, content-curation, bookmarking, article-retrieval, knowledge-base, search-indexing ## Description This Connector connects your Omnivore account to your AI agent so you can organize your reading list and pull knowledge from saved articles using natural language. Instead of digging through folders or remembering which label you used for a specific research paper, you can just ask your agent to find what you need. You can pull the full text of a saved piece into your current conversation to get a summary, analyze a specific section, or cross-reference multiple sources at once. When you stumble upon something new while browsing, you can have your agent save it to your library instantly. It turns a static list of bookmarks into a dynamic research assistant. By connecting this through the Vinkius catalog, you're essentially giving your agent a memory of everything you've ever wanted to learn. You stop wasting time on manual searching and start spending it on actually processing the information you've already collected. ## Tools ### search_articles Finds specific content using labels, folders, or read status filters like is:unread. ### get_article Fetches the complete text and metadata for a specific article so your agent can read the whole thing. ### get_me Retrieves your current account information to verify that your connection is active. ### save_url Adds a new web link to your Omnivore library immediately from your chat window. ## Prompt Examples **Prompt:** ``` Search my Omnivore library for unread articles about 'Machine Learning'. ``` **Response:** ``` I found 3 unread articles about Machine Learning. The most recent is **Understanding Transformers** (slug: understanding-transformers). Would you like me to fetch the full content of any of these? ``` **Prompt:** ``` Fetch the full content of the article with slug 'mcp-guide' for username 'alex_dev'. ``` **Response:** ``` I've retrieved the article **The Ultimate MCP Guide**. It covers: * Model Context Protocol architecture * Implementation details * Best practices Would you like a summary of the key points? ``` **Prompt:** ``` Save the URL 'https://blog.omnivore.app/p/getting-started' to my library. ``` **Response:** ``` Successfully saved the URL to your Omnivore library. It is now available in your inbox for later reading. ``` ## Capabilities ### Find specific articles Search your saved content using labels, folders, or read status filters. ### Pull full article text Fetch the complete content of a saved link for deep analysis or summarization. ### Save new links instantly Add new web links to your library directly from your conversation. ### Check account status Verify your connection and profile details with your agent. ### Filter unread content Quickly isolate articles you haven't looked at yet in specific categories. ### Extract author info Get metadata like author names and labels for any saved piece. ## Use Cases ### Finding a specific paper A researcher asks for unread articles about machine learning and gets a list of links to choose from. ### Summarizing a long read A student provides a link and asks the agent to save it and then summarize the main points in a few bullets. ### Content research A blogger asks the agent to find all articles labeled newsletter to find inspiration for a new post. ### Quick bookmarking A user finds a great blog post and tells the agent to save it to their Omnivore library for later. ## Benefits - Stop manual searching by using search_articles to find content by label or folder instantly. - Get full context by using get_article to pull complete text into your chat for deep analysis. - Save time by using save_url to bookmark new links without leaving your current AI conversation. - Organize your research better by asking your agent to filter your unread library. - Access your entire reading history through a single interface in Claude, Cursor, or Windsurf. ## How It Works The bottom line is it turns your saved bookmarks into an interactive knowledge base your AI can actually use. 1. Subscribe to the Omnivore (Read-Later) MCP and provide your API key. 2. Connect the Connector to your preferred AI client like Claude or Cursor. 3. Ask your agent to search, save, or read your content. ## Frequently Asked Questions **Can I use Omnivore (Read-Later) MCP to find my saved bookmarks?** Yes. It lets your AI agent search through your entire Omnivore library using labels, folders, and read status to find exactly what you're looking for. **How does Omnivore (Read-Later) MCP help with research?** It lets your agent pull the full text of saved articles into your chat. This makes it easy to summarize long pieces or find specific information across multiple saved sources quickly. **Can I save new links to my Omnivore library using an AI agent?** Yes. You can just give your agent a URL and tell it to save it. It will then add that link to your Omnivore library automatically so you can read it later. **Does Omnivore (Read-Later) MCP work with Claude or Cursor?** Yes. It works with any MCP-compatible client, including Claude, Cursor, and Windsurf, giving those tools access to your saved content. **Can I use Omnivore (Read-Later) MCP to see my account details?** Yes. You can ask your agent to check your account info to verify your connection status and ensure your API key is working correctly. **Will Omnivore (Read-Later) MCP help me organize my reading list?** It makes your list much more useful by allowing your agent to filter content. You can ask it to find only unread articles or content within specific folders. **Can I filter my search by labels or read status?** Yes. Use the `search_articles` tool with Omnivore's search syntax, such as `label:AI` or `is:unread`, to narrow down your results. **How do I get the actual text of a saved page for analysis?** Use the `get_article` tool by providing the article's unique slug and the owner's username. The agent will retrieve the full text content and metadata. **Is it possible to add new links to my library via the agent?** Yes, the `save_url` action allows you to send any web link directly to your Omnivore library for later reading.