Ollama logo Ollama logo background glow

Ollama

A command-line tool and REST API for running and managing open-weight large language models on local hardware or a hosted cloud service

It takes care to curate a safe software space. Help us do more.

&

+
Local Model Execution
local, free, pro, max, team, enterprise
Runs open-weight large language models entirely on the user’s own CPU or GPU, allowing offline use and keeping prompts and outputs off any external server.
+
REST API
local, free, pro, max, team, enterprise
Exposes a local REST API on port 11434 for chat, generation, and embeddings, enabling developers to integrate model inference into their own scripts and applications without extra setup.
+
Modelfile Customization
local, free, pro, max, team, enterprise
Lets users define custom model configurations, system prompts, and parameters through a Modelfile, allowing them to create and share tailored model variants.
+
Tool/Function Calling
local, free, pro, max, team, enterprise
Lets supported models invoke external functions mid-conversation, enabling agentic workflows such as calculators, web search, or custom application logic.
+
Structured Outputs
local, free, pro, max, team, enterprise
Constrains a model’s response to a JSON schema defined by the user, so applications can reliably parse machine-readable output without extra prompt engineering.
+
Multimodal Model Support
local, free, pro, max, team, enterprise
Runs vision-capable models that accept image input alongside text, allowing users to analyze or describe images without sending them to a cloud service.
+
OpenAI-Compatible Endpoint
local, free, pro, max, team, enterprise
Mirrors the OpenAI Chat Completions API format, letting existing OpenAI-client code work against local models with minimal code changes.
+
Coding-Agent Integrations
local, free, pro, max, team, enterprise
Provides one-command launchers for tools such as Claude Code, Codex, and OpenCode, cutting the setup time needed to point a coding agent at a local or cloud model.
+
Cloud Model Offloading
free, pro, max, team, enterprise
Runs larger models on Ollama’s hosted infrastructure while keeping the same local CLI and API workflow, so users can access models too large for their own hardware.
...7 more features/limitations. Contact us to get a complete list of features and system requirements.

Platform

Desktop
Software

Social

System Requirements

#Minimum
1
Windows 10 22H2 or newer, Home or Pro
2
  • NVIDIA GPU
    • with compute capability 5.0+ and driver 550 or newer
    • driver 570+ required for compute capability 5.0-6.2
  • AMD Radeon GPU with ROCm v7 / HIP7-capable driver
#Minimum
1
macOS 14 Sonoma or newer
2
Apple Silicon (CPU and GPU support) or x86-64 (CPU only)
#Minimum
1
  • NVIDIA GPU
    • with compute capability 5.0+ and driver 550 or newer
    • driver 570+ required for compute capability 5.0-6.2)
  • AMD Radeon GPU with ROCm v7 driver
Want to get access to entire data for training LLM or studying software or may be something else? Contact us at contact [at] softorage [dot] com

Ratings

4.30
5

G2
4.3
5
based on 5 reviews

Developer

Written in

Go

Initial Release

2023-07-07

Repository

License

Categories


Notes

  • Running models locally is always free and unlimited; the Free/Pro/Max/Team/Enterprise tiers govern usage of Ollama’s hosted cloud models only, not local/self-hosted use.
  • Beginner’s guide: Ollama quickstart