An open-source tool to download and run open large language models locally with one command, exposing an OpenAI-compatible API.
Ollama pulls and runs open models (Llama, Mistral, Gemma, Qwen, and many more) on your own machine — `ollama run llama3` — and serves them over a local HTTP API that mirrors the OpenAI format, so existing tools and code point at it with a URL change. It handles model files, quantisation, and GPU offload for you, and integrates with editors and agent frameworks like Claude Code, VS Code, and n8n.
Why it's a discovery: it's the simplest path to private, offline LLM inference, used by millions of developers. The local software is free and open source; there's an optional paid cloud for larger models.
Best for: developers prototyping with LLMs without API bills, and anyone who needs prompts to stay on their own hardware.
This hasn't been featured in a video yet.
