# OneAI MCP for AI Agents AI Agent Connect > OneAI lets your AI agent handle heavy lifting for text, audio, and video analysis. It turns long documents, podcast files, and customer feedback into structured summaries, sentiment scores, and extracted entities. Instead of manual transcription or reading through hundreds of emails, you can just ask your agent to process the data and give you the highlights. ## Overview - **Category:** developer-tools - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_OUuZaJpMXxYZjYxd7ELA63wi1sMTV0hm1c50LuC2/ai-agent-connect - **Tags:** nlp, transcription, summarization, text-analysis, ai-skills ## Description OneAI gives your AI agent a set of specialized skills to handle media and text that usually require a lot of manual effort. If you are dealing with hours of podcast audio, long research papers, or a mountain of customer reviews, this Connector handles the heavy lifting. It does not just read text; it understands context to pull out specific entities, gauge sentiment, and create concise summaries. You can run these tasks as quick, synchronous requests or set up asynchronous pipelines for massive files that need more time to process. Because it is hosted on the Vinkius Marketplace, you can connect it to your existing setup and start using these language skills immediately without worrying about building your own NLP models. It is about getting straight to the insights without the tedious work of transcription or manual data sorting. Instead of spending your afternoon listening to a recording to find one quote, you can have your agent do it in seconds. You can also feed it a long URL and get a summary of the key points without clicking through multiple pages. This tool bridges the gap between raw data and actionable information, whether that data is trapped in a video file, a PDF, or a messy spreadsheet of feedback. It works by applying pre-trained language models to your specific inputs, ensuring that the output is structured and ready for your next step. ## Tools ### get_async_task_status Check if a long-running file process is finished and grab the data. Use this to see when your transcription or large file analysis is ready. ### run_async_pipeline Start a background job for large files or audio transcription. It is the best way to handle media that takes a few minutes to process. ### run_pipeline Run a quick summary, sentiment check, or entity extraction on a piece of text. This allows you to get immediate results for short content. ## Prompt Examples **Prompt:** ``` Summarize this article and extract entities: [URL] ``` **Response:** ``` I've processed the article for you. Here is the summary: **Key Summary** * The study highlights a 20% increase in renewable energy adoption. * New policies in the EU are driving this shift. * Infrastructure costs remain the primary hurdle. **Extracted Entities** * **Organizations:** European Commission, Global Energy Corp * **Locations:** Germany, Spain, Denmark * **Dates:** 2023-2025 ``` **Prompt:** ``` Transcribe the audio at this URL: https://example.com/podcast.mp3 ``` **Response:** ``` I've started the transcription for that podcast. **Status:** Processing **Task ID:** `task_8821` I'll let you know as soon as the text is ready, or you can ask me to check the status anytime. ``` **Prompt:** ``` Check the status of task task_8821. ``` **Response:** ``` The transcription for your podcast is complete. Here is the text: "Welcome back to the show. Today we're discussing the shift in renewable energy... [Full transcript follows]" ``` ## Capabilities ### Summarize long articles Turn long-form content into concise bullet points or paragraphs. ### Transcribe audio files Convert spoken word from audio or video into readable text. ### Extract entities from text Identify and pull out names, dates, and other key data points. ### Detect sentiment in feedback Automatically determine the emotional tone of customer reviews. ### Process large video files Handle high-volume media files using background processing. ### Track long-running tasks Monitor the progress of large data analysis jobs in real time. ## Use Cases ### Podcast Repurposing A podcaster asks their agent to transcribe a 90-minute interview and generate a summary of the key takeaways. ### Customer Feedback Loop An analyst asks the agent to scan 500 reviews and extract a list of common product complaints and sentiments. ### Research Paper Extraction A researcher provides a URL to a long white paper and asks the agent to extract all mentioned dates and locations. ### Video Content Summarization A creator gives the agent a video link and asks for a 3-sentence summary of the main points discussed. ## Benefits - Save hours on transcription by converting long podcast audio into clean text. - Get organized data from messy feedback by automatically pulling out sentiment and key entities. - Handle massive documents without hitting limits by using background processing for large files. - Skip the model training headache by using pre-trained language skills for common NLP tasks. - Track progress on big jobs automatically to see when your data is ready for use. ## How It Works The bottom line is it turns raw media and text into organized data points in a few clicks. 1. Subscribe to the OneAI MCP and grab your API key. 2. Enter your API key into your AI client configuration. 3. Ask your agent to summarize a link or transcribe a file to see the results. ## Frequently Asked Questions **Can OneAI transcribe my podcast files?** Yes. OneAI can convert audio and video files into text, making it easy to turn your spoken content into blog posts or show notes. **How does OneAI handle very long documents?** OneAI uses background processing for large files. This means it can handle long research papers or massive documents that would normally overwhelm a standard chat. **Can OneAI tell me if customer reviews are positive or negative?** Yes. It includes sentiment analysis skills that allow your agent to gauge the emotional tone of feedback and summarize the main complaints. **Do I need to build my own models to use OneAI?** No. OneAI comes with pre-trained language skills, so you can start analyzing text and media immediately without any technical setup. **Does OneAI work with Claude or Cursor?** Yes. It is a standard MCP that connects to any compatible client, including Claude, Cursor, and Windsurf. **Can I extract specific names and dates from a text block?** OneAI can identify and pull out specific entities like names, dates, and locations from any text you provide. **How can I summarize a text and extract entities at the same time?** Use the `run_pipeline` tool. Provide your text in the `input` field and define the `steps` as a JSON array like `[{"skill":"summarize"}, {"skill":"entities"}]`. **What should I use for processing large audio files?** Use `run_async_pipeline`. You can provide a `content_url` for the audio file and set the `steps` to include the `transcribe` skill. This starts a background task. **How do I know when my asynchronous processing is finished?** Use the `get_async_task_status` tool with the `task_id` returned by the async pipeline. It will provide the current status and the final results once completed.