ollama/ollama-js

Ollama JavaScript Library lets you add AI chat features to your JavaScript or TypeScript app without dealing with complex API plumbing. Instead of writing raw network requests to talk to an AI model, you install this package, write a few lines of code, and you can send prompts to open-source models like Llama and get text back. It works in both Node.js server environments and in the browser. At its core, the library is a friendly wrapper around the Ollama REST API. You call a method like chat, pass it a model name and a list of messages, and you get a response. If you want the answer to appear word-by-word as it's generated, you flip on streaming and the library hands you each piece as it arrives. Beyond chatting, it covers the full lifecycle of working with models, you can pull new ones, list what you have, delete, copy, and even create custom versions. It also supports generating text from a single prompt, creating embeddings (numerical representations of text useful for search), and sending images to vision-capable models. This is for developers building apps who want to run AI models locally or through Ollama's cloud service. A founder prototyping a support bot could use it to test different models on their laptop before committing to a paid API. A PM building an internal tool could add a feature that summarizes documents or answers questions from a knowledge base. The library also supports cloud models, you can sign in, pull a larger model that wouldn't fit on your machine, and the library transparently offloads the work to Ollama's cloud. A few things stand out. The library lets you mix local and cloud models with the same code, just by changing the model name or pointing the client at a different host. It also includes web search and web fetch capabilities, though those require an Ollama account. The API supports advanced features like tool calling, structured JSON output, and model "thinking" modes, but none of that is forced on you, the simplest usage is genuinely just a few lines.