Ollama
MIT-licensed local model runner for macOS, Windows and Linux that downloads open models and serves them from one command, with an OpenAI-compatible API on the machine it runs on.
What it does
- Model inference
- Model hosting
Features 14
- StreamingRenders text as the model produces it; on by default in the REST API.
- ThinkingThinking-capable models emit a separate field carrying their reasoning trace.
- Structured outputsEnforces a JSON schema on model responses for reliable structured data.
- VisionVision models accept images alongside text to describe, classify and answer questions.
- EmbeddingsGenerates text embeddings for semantic search, retrieval and RAG pipelines.
- Tool callingLets a model invoke tools (function calling) and incorporate the results into its replies.
- Web searchA web search API that augments models with the latest information.
- SpeechSpeech models read text aloud, and Ollama streams the audio as it is generated.
- Web fetch APIFetches a single web page by URL and returns its content.
- Decision (System One)Use System One to classify text, answer yes/no questions, or score text against a rubric.
- ModelfileA Modelfile is the blueprint to create and share customized models using Ollama.
- ollama launchConfigure and launch external applications to use Ollama models.
- OpenAI compatibilityConnect OpenAI clients to Ollama, which supports a subset of the OpenAI API.
- Anthropic compatibilityConnect Anthropic clients and tools such as Claude Code to Ollama.
How it compares
Sources
- Pricing
-
Official documentation · 18 Sep 2026
Licence read from the GitHub API's own license field: MIT.
- Description
-
Official documentation · 18 Sep 2026
Homepage, fetched 2026-09-18: 'Nothing you run locally ever leaves your machine.'
- Sold within
-
Official documentation · 18 Sep 2026
Downloaded and run on its own for macOS, Linux and Windows; no parent offering is involved.
- URL
-
Official documentation · 18 Sep 2026
Fetched 2026-09-18, 200.
- Interfaces
-
Official documentation · 12 Oct 2026
Page: 'for macOS, Windows, or Linux. Open the app, or get started from your terminal'.
- Modalities
-
Official documentation · 1 Oct 2026
Page: 'Vision models accept images alongside text so the model can describe, classify, and answer questions about what it sees'.
- Intended for
-
Official documentation · 13 Oct 2026
Page: 'Ollama is the most popular way to build with open models'.
Also from Ollama
-
Ollama Cloud Model API
Hosted inference for the same open models, called from the Ollama CLI or an application with an API key and served from the US, Europe and Singapore.
Something wrong here? Send a correction — quote this product id: ollama.