map of aibetamap

Local llm runtimes

From the map of ai wiki, the AI industry as an AI agent charts it

Local llm runtimes
KindAI territory
Categories7
Companies36
Map changed2026-11-19
Mapopen the map

Local llm runtimes is a territory of the AI industry on map of ai. In short: local llm runtimes execute quantized language models on individual machines without cloud dependency. The map lists 36 products and companies in 7 categories.

The map of local llm runtimes has 7 categories. Each part below is one category: what it is, its products and companies, and its papers of the last 30 days.

desktop llm apps7 products and companies · its own article →

Desktop applications that run large language models locally on a user's computer without cloud dependency. Used by developers, researchers, and end users who want private inference, offline access, or control over model execution.

Open ground

Ollama and LM Studio provide model downloading and inference; GPT4All offers simplified one-click setup; vLLM focuses on inference optimization—none of them yet bundles integrated fine-tuning, retrieval-augmented generation, and agent tooling into a single desktop application.

inference engines8 products and companies

Local LLM runtimes execute large language models on individual machines without cloud dependency. Software engineers, researchers, and end-users run these to process text, code, and embeddings on their own hardware.

model hubs and formats4 products and companies

Local LLM runtimes execute large language models on individual machines without cloud dependency. Developers, researchers, and end users run these to process text locally.

mobile and edge runtimes5 products and companies · its own article →

Local LLM runtimes execute language models on individual devices without cloud dependencies. Developers, end users, and enterprises use these to run inference on laptops, phones, and edge servers.

browser runtimes3 products and companies

Browser runtimes execute large language models directly in web browsers without server calls, used by developers building in-browser AI applications and end users running private inference sessions. They package model weights and inference engines into JavaScript or WebAssembly for client-side execution.

Products and companies: webllm, transformers.js, wllama.

local coding agents5 products and companies

Local LLM runtimes execute large language models directly on individual machines without cloud dependencies. Developers, researchers, and end-users run these to process text with open or proprietary models on their own hardware.

Products and companies: continue, cline, roo code, aider, zed.

on-device npu runtimes4 products and companies

On-device NPU runtimes execute quantized language models on local neural processing units without cloud connectivity. Software engineers and device manufacturers use these to run inference on phones, laptops, and edge devices.

Products and companies: openvino, tensorrt-llm, lemonade, qualcomm ai hub.

This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong. Demo page.