Ollama

What Ollama does

Ollama is an open-source platform that simplifies building with open models. It allows users to run AI models locally on their hardware or access larger, more powerful models in Ollama's cloud. The platform offers CLI, API, and desktop apps, along with extensive community integrations. It supports various tasks like coding automation, document analysis, and general AI assistance, with features for private data handling and scalable cloud usage plans.

Key capabilities

• Run models locally on your hardware
• Access cloud models
• CLI, API, and desktop apps
• 40,000+ community integrations
• Unlimited public models
• Access larger, more powerful cloud models
• Upload and share private models
• Shared usage across your team (Team plan)
• Centralized billing and administration (Team plan)
• Single sign-on (SSO) (Team plan)
• Model access controls (Team plan)
• MDM installer for Windows and macOS (Team plan)
• Priority support and dedicated Slack channel (Team plan)
• Tool calling support for cloud models
• Native weights for cloud models
• Accelerated data formats for modern NVIDIA hardware
• Low time-to-first-token and high throughput
• Usage measured by actual GPU time, not fixed tokens
• Concurrency limits for running multiple models simultaneously
• Python library
• JavaScript library

Use cases

• Automate coding
• Document analysis
• Chatting with models
• Evaluating larger models
• Coding and AI assistants with smaller models
• Larger models for day-to-day work
• Coding automation
• Deep research
• Continuous agent tasks
• Multiple concurrent agents
• Large models over extended sessions
• Coding agents
• Personal assistants
• Editors
• Chat
• Vision
• Embeddings
• Reasoning

AI approach

Ollama provides an open-source platform for running AI models locally and offers access to larger models via its cloud. Users can automate tasks, analyze documents, and use models for chat, coding, vision, embeddings, and reasoning. The platform supports tool calling for cloud models and focuses on low time-to-first-token and high throughput.

Tech named: NVIDIA hardware, Blackwell architecture, Vera Rubin architecture, NVFP4

What it says sets it apart

• Open-source platform
• Ability to run models locally for data privacy
• Flexible cloud model access with varying usage tiers
• Usage measured by GPU time, not tokens, adapting to hardware efficiency
• Extensive community integrations
• Support for tool calling in cloud models
• Dedicated capacity for workflows needing multiple models concurrently

Funding rounds we track

From the AI funding tracker — rounds as reported by the linked publications.