- AI
- Software that performs tasks such as recognising patterns, generating text, or making predictions.
- AI agent
- An AI system that can choose steps and use tools to work on a task.
- API
- A way for one app to request data or actions from another app.
- Chatbot
- An app you interact with by sending messages.
- Context window
- The amount of information a model can consider in a single request.
- Embedding
- A numerical representation used to find similar text, images, or other data.
- Fine-tuning
- Training an existing model further on examples for a particular task.
- Free
- The listed option has no software fee. Other costs, such as hosting, may apply.
- Freemium
- A free plan with limits, plus optional paid features or higher usage.
- Free trial
- Temporary access that may require payment after the trial ends.
- Generative AI
- AI that creates content, such as text, images, audio, or code.
- GPU
- A processor often used to run or train AI models efficiently.
- Hallucination
- An answer generated by a model that sounds convincing but is incorrect or unsupported.
- Inference
- Running a trained model to produce an answer or prediction.
- LLM
- A large language model: a model trained to work with language.
- Local AI
- AI software or models that run on your own device.
- Multimodal
- Able to work with more than one kind of input, such as text and images.
- Open source
- Software distributed under a license that permits access to its source code and specified uses.
- Paid
- A plan or purchase is required for the listed option.
- Prompt
- The instructions and information you give an AI model.
- RAG
- Retrieval augmented generation: finding relevant information before asking a model to answer.
- Self-hosted
- Software you install and operate on a device or server you control.
- Token
- A piece of text a model processes. Usage charges often depend on token counts.
- Workflow
- A sequence of steps that turns an input into a result.
