Ollama

What it does

Ollama takes prompts and requests for open models and returns model responses. Models can run on your own computer, or be called as cloud models through Ollama's API and compatible OpenAI and Anthropic clients. Cloud models are billed by token at per-model rates, while local runs use your own hardware.

People use Ollama to download and run models on their desktop, connect desktop apps such as Claude or ChatGPT, launch coding agents from a terminal, or build models into their own applications through its Python and JavaScript libraries. Its distinct feature is combining local and hosted use of the same open model library under one account, with a stated policy that prompts and responses are never logged or trained on.

Ollama suits developers, hobbyists and teams who want open models and control over where inference runs. Local use is free to run on your own hardware, while cloud usage requires an Ollama account and draws from plan credits or pay-as-you-go balance. Cloud concurrency is limited by plan, and Enterprise pricing is custom.

What it can do

• Download and run open models on your own computer
• Run cloud-hosted open models without a local GPU
• Call models through Ollama's API or compatible OpenAI and Anthropic clients
• Use Python and JavaScript libraries to build models into applications
• Connect desktop apps and launch coding agents from the terminal
• Browse a model library for coding, vision, embeddings and reasoning

Works with

• OpenAI-compatible API clients
• Anthropic-compatible API clients
• Claude and ChatGPT desktop apps (via integrations)

Example workflows

Run a local model for private drafting

Input: a text prompt typed into Ollama's local interface. Steps: download a model from the library, run it on your computer. Output: a response generated entirely on your hardware.

Build a cloud-model app

Input: an app request sent to the Ollama API or Python/JavaScript library. Steps: create an account, get an API key, call a cloud model. Output: a model response returned to your application, billed per token.

Before you choose

• Cloud usage is billed by token against included plan credits or pay-as-you-go balance, so costs depend on model and volume.
• Concurrent cloud requests are capped by plan (3 for Pay as you go and Pro, 10 for Max and Team), with excess requests queued.

Listed under

• Coding
• Sandboxes and infrastructure
• API
• Cloud inference
• Developer tools
• LLM runtime
• Local inference
• Open models

Get started

Download the app and run a model locally, or sign up for cloud access.

1. Download Ollama from the download page and install it on your computer.
2. Download and run a model from the model library, locally or via a cloud model.
3. Sign in to an Ollama account to use cloud models and choose a plan or pay-as-you-go.

Running models on your own hardware is always unlimited. Cloud usage is billed per million tokens at each model's rate, with input, cached input and output priced separately and off-peak discounts. Pro and Max annual billing is shown per month. Team requires a monthly subscription and shares usage credits across the team. Check the model pricing table for current per-model rates.

Source: ollama.com/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Ollama · Run AI models locally and in the cloud ↗ollama.com
• Ollama Pricing ↗ollama.com
• Ollama documentation - Ollama ↗docs.ollama.com
• Download Ollama on macOS ↗ollama.com

Company website

ollama.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.