> For the complete documentation index, see [llms.txt](https://deeptrain.gitbook.io/documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://deeptrain.gitbook.io/documentation/getting-started/introduction.md). # Introduction
## What is Deeptrain? Deeptrain is a multi-modal data connector for LLMs and AI agents. We help you source and integrate data that is not directly available and understandable by transformer models and AI. ## Multi-modal Capabilities Deeptrain helps you build truly mulimodal applications where your AI agents can learn and interact with content that is beyond the processing capabilities of currently existing LLMs. ### Text Train and customise your AI agents beyond the pre-existing context window size using localized embedding database which retrieves content in real-time from live data sources to back your AI's answers. ### Images Easily turn any non-vision supported model into a computer-vision integrated LLM using Deeptrain. ### Flowchart, graphs Help AI understand flowcharts, graphs, diagrams and more using Deeptrain integration with your AI agents and models. ### Audio Proces any audio content and use it for training and processing for your AI agents. ### Videos NEW : Deeptrain now supports multi-dimensional video processing. You can now use an video, hosted locally, self-hosted, or on supporting platform like Vimeo or YouTube to process them and use it as your AI's agent knowledge. You can take videos from your users as input filed using [Deeptrain's Transcribe API. ](/documentation/getting-started/transcribe-api.md) ## Multi-dimensional Capabilities Deeptrain is not limited to one particular language model. In fact, you ca equip 200+ private or open-source models with Deeptrain's tech & services. Deeptrain will soon be able to process custom models too.