LeMUR is a powerful framework that enables developers to build generative AI applications on audio data with ease. It seamlessly integrates all the essential components, including automatic transcription, prompt augmentation, compression, retrieval techniques, language models, and structured outputs, into a single API.

With LeMUR, you can effortlessly search, summarize, ask questions, or generate new text based on the knowledge of your application's spoken data. This allows you to create AI-powered apps for meetings, phone calls, videos, podcasts, and more.

LeMUR is designed to be highly accurate on core tasks such as summarization, question-answering, and action item generation. It can also be extended to any other use case with a customizable endpoint, giving you the flexibility to define your own tasks and prompts.

To get started, you can use the Assembly AI Python SDK or any other programming language. The LeMUR playground allows you to upload files or enter YouTube links and experiment with all the available endpoints for free. Additionally, you can sign up for a free API token and explore the comprehensive documentation and welcome Colab to get up and running quickly.

LeMUR's powerful transcription and summarization capabilities allow you to effortlessly process and summarize large volumes of audio data. With the ability to ingest over 1 million tokens, equivalent to approximately 100 hours of audio, LeMUR enables you to summarize entire lecture series, virtual meetings, phone calls, podcasts, and more.

To further enhance the summarization process, LeMUR allows you to provide additional context in the form of text, enabling the model to pay particular attention to specific topics or areas of interest. This ensures that the generated summaries are tailored to your needs, providing you with the most relevant and informative insights.

Whether you're working with customer service call recordings, educational content, or any other type of spoken data, LeMUR's summarization capabilities empower you to quickly and efficiently extract the key information, saving you time and effort.

LeMUR's Question-Answering (Q&A) endpoint allows you to gain deep insights from your spoken data. You can ask questions about the content of your virtual meetings, phone calls, podcasts, and more, and LeMUR will provide accurate and contextual answers.

Whether you need to understand a customer's history in a call center or clarify a concept mentioned in a podcast, LeMUR's Q&A capabilities can help you find the information you need. The answers provided will include relevant reasoning and citations, ensuring you have a comprehensive understanding of the spoken data.

LeMUR's Q&A is designed to be highly flexible, allowing you to customize the prompts and tasks to fit your specific use cases. This makes it a powerful tool for unlocking the value of your spoken data and driving informed decision-making.

LeMUR's Action Items endpoint allows you to automatically generate a list of action items from virtual meetings. You can provide a specific format to follow and add context on the speakers to assign action items to specific meeting attendees. This feature helps streamline meeting productivity by ensuring that all important tasks and responsibilities are clearly identified and assigned, enabling teams to stay organized and accountable. With LeMUR, you can effortlessly extract actionable insights from your meeting recordings, freeing up time and resources for more strategic initiatives.

LeMUR's flexible and extensible design allows you to unlock the full potential of your spoken data. The customizable endpoint empowers you to define your own unique tasks and prompts, enabling you to build AI applications tailored to your specific needs.

Whether you're looking to extract insights from customer service calls, generate summaries for podcasts, or develop novel applications that leverage the knowledge embedded in your audio data, LeMUR's customizable endpoint provides the tools you need to bring your ideas to life.

By leveraging the power of large language models and LeMUR's robust processing capabilities, you can seamlessly integrate spoken data into your workflows and unlock new possibilities for innovation. Explore the boundaries of what's possible and let your creativity shine through as you build transformative audio-driven applications with LeMUR's customizable endpoint.

The easiest way to get started with LeMUR is through the playground, where you can upload a file or enter a YouTube link and experiment with all LeMUR endpoints with just a few clicks, for free. To get started, simply sign up using the link in the description and obtain a free API token.

If you prefer to use the API directly, our comprehensive documentation and a welcome Colab notebook will have you up and running in just a few minutes. LeMUR is designed to be highly accurate on a core set of tasks, including summarization, question-answering, and action item generation, and can also be extended to any other use case with a customizable endpoint.

We're excited to see what you'll build with LeMUR, so please share your thoughts and projects in the comments below. The LeMUR team is eager to hear your feedback and support your journey in leveraging the power of large language models for your audio-based applications.

Lemur is a powerful framework that simplifies the process of applying large language models (LLMs) to transcribed speech. It provides a comprehensive set of tools, including automatic transcription, prompt augmentation, compression, and retrieval techniques, as well as language models and structured outputs, all accessible through a single API.

With Lemur, developers can easily build AI applications that can search, summarize, ask questions, and generate new text based on their spoken data, such as virtual meetings, phone calls, videos, and podcasts. The framework's ability to handle large amounts of audio data, up to 100 hours, and its customizable endpoints make it a versatile solution for a wide range of use cases.

Lemur is now available for everyone to use, with new endpoints, higher accuracy outputs, and higher input and output limits. The easiest way to get started is through the Assembly AI Python SDK or the Lemur playground, where you can experiment with the various endpoints for free. The Lemur documentation and welcome Colab provide additional resources to help you get up and running quickly.

We're excited to see what you'll build with Lemur and encourage you to share your thoughts and feedback in the comments below.