Home / Glossary / Local AI models and agents
Glossary
What are local AI models and local AI agents?
Local AI models and local AI agents run on a person's own device instead of a vendor's cloud. A local model answers prompts using the machine's own power. A local agent goes further: it can read files, run commands and call tools. Because they run on the device, they may never touch a gateway.
By the AIBound team · Updated 15 November 2026
Key takeaways
- Local AI runs on the employee's device, not in a vendor's cloud.
- A local model answers prompts, and a local agent can also take actions.
- Gateways may never see local AI, so security has to look at the device.
Why local AI matters
A local model can hold company data on the device, and a local agent can act on it. Neither needs to send anything through a gateway to do its work. That makes local AI easy to miss, and it can run for months before anyone asks what is on the laptop.
How local AI gets onto a device
Local AI costs nothing extra to run once it is downloaded, which is why it spreads quickly.
- 1 An employee downloads a local model or installs a local agent.
- 2 The model or agent runs on the device, using its files and settings.
- 3 Nothing has to leave the device, so no gateway or cloud tool records it.
Local AI model vs local AI agent
The two are often installed together, but they do different jobs.
| Question | Local AI model | Local AI agent |
|---|---|---|
| What is it? | A model that runs on the device and answers prompts | A program that uses a model to take actions |
| What can it do? | Produce text, code or answers | Read files, run commands and call tools |
| Why does security care? | It may hold company data on the device | It can act on that data and reach systems |
Examples
Examples of local AI
FAQ
Frequently asked questions
What is a local AI model?
A local AI model is an AI model that runs on a person's own device instead of a vendor's cloud. You download it, and it answers prompts using the machine's own power. It can work without an internet connection, so it may leave no trace on the network.
What is a local AI agent?
A local AI agent is a program on a device that uses an AI model to take actions. It can read files, run commands and call tools, often with the permissions of the person who installed it. That reach is why a local agent needs more attention than a model that only answers.
Why do local AI models matter for security?
They run where most security tools do not look. A local model can hold company data on the device, and a local agent can act on it, with no gateway log either way. Security teams usually learn about them late, if at all, unless they look at the device.
How do you find local AI on employee devices?
Look at the device. AIBound installs on employee devices in 15 minutes, pushed by IT like any other app, and finds local models, agents, MCP servers and skills in real time, with the person who runs each one. You then choose what to allow, ask about or deny.