L

An open-source Python and JavaScript framework for building applications powered by large language models. LangChain provides abstractions for chaining LLM calls, connecting to external data sources, building memory systems, and creating AI agents. It is one of the most widely used frameworks for production AI application development.

A type of AI model trained on massive amounts of text data that can understand and generate human language. LLMs are the foundation of tools like ChatGPT, Claude, and Gemini. They work by learning statistical patterns in language and using those patterns to predict and generate relevant, coherent text in response to a prompt.

The time delay between sending a request to an AI model and receiving a response. Low latency is critical for real-time applications like AI chatbots and voice assistants. Latency is influenced by model size, server load, and network speed.

The internal mathematical representation space where an AI model encodes the meaning of inputs. In a latent space, similar concepts are positioned close together — so 'king' minus 'man' plus 'woman' equals 'queen' in a word embedding space. Latent space is the foundation of how AI models understand relationships between concepts, and it is the space in which image generation models like Stable Diffusion operate.

Meta AI's open-source large language model family, widely used by researchers and businesses who want to run AI models locally or on private infrastructure. Llama 4 (released April 2025) includes Scout (17B parameters, 10M token context) and Maverick (400B parameters) variants. Llama models can be fine-tuned and deployed without per-token API costs.

An open-source data framework for building LLM-powered applications that need to ingest, index, and query large amounts of custom data. LlamaIndex specializes in connecting LLMs to external data sources — documents, databases, APIs — and is particularly strong for building RAG (Retrieval-Augmented Generation) pipelines. It is a popular alternative to LangChain for data-heavy AI applications.

A proposed standard file (placed at yourdomain.com/llms.txt) that provides AI crawlers with a structured, machine-readable summary of your website's content, purpose, and key facts. Similar to robots.txt for search engines, llms.txt helps AI systems understand and accurately represent your content when generating answers. Fewer than 1% of websites have implemented llms.txt as of 2026.

An AI-powered full-stack web application builder that generates complete, deployable web apps from natural language descriptions. Lovable is a leading no-code/vibe coding platform for building SaaS products, internal tools, and landing pages without traditional software development skills.

An AI company known for two flagship products: Dream Machine (text-to-video and image-to-video generation) and NeRF-based 3D scene capture. Dream Machine generates smooth, high-quality video clips from text prompts and is widely used by filmmakers and marketers. Luma's 3D capture technology lets users create photorealistic 3D models of real spaces using only a smartphone.