AI App Catalog · Local LLM

llama.cpp at Work: Risk and Controls

ggml · Local LLM. What an MSP needs to know to govern llama.cpp on client devices.

At a Glance

Vendorggml
Websitegithub.com/ggml-org/llama.cpp
CategoryLocal LLM
Default Risk RatingMedium
Default Status for New ClientsUnapproved
Recognized ByIts installed app, found by the device inventory
Desktop App InventoriedmacOS, Linux
Extension InventoriedVS Code, Cursor and Windsurf
CoverageInventory
Catalog ReviewedOctober 28, 2026

The risk rating and default status are ProctoredAI's starting points for a new client. Each MSP sets the status per client. How apps are rated and verified.

What to Know About llama.cpp

Local model runtime and server (port 8080 by default); models and chats stay on the device.

Why Local LLM Tools Need Governing

Local runtimes run models on the device, so nothing is inspected on the way out, and downloaded models can be anything.

Inventory the runtimes and models on every device and decide per client whether local models are allowed.

How ProctoredAI Governs llama.cpp

  • See it. The agent reports every device with llama.cpp installed, and finds its extension in the browser or code editor.
  • Decide it. Mark it Approved, Unapproved or Blocked for each client. Unapproved use can ask the person for a business reason first.
  • Protect what is sent. ProctoredAI finds the installed app; set it Blocked or Unapproved per client until a use case is approved.
  • Prove it. Usage, decisions and blocked events land in the client's monthly report as evidence.

Policy Packs That Fit

  • Credentials & Secrets. Block API keys, private keys, passwords and connection strings everywhere.
  • Sensitive Documents. Contracts, financial statements, HR and medical records, customer lists and credential files may only go to approved AI apps.

Other Local LLM Apps

All 1225 AI apps in the catalog · What ProctoredAI does

Run AI Governance for Every Client From One Console

Deploy one agent from your RMM, choose what each client may use, and stop sensitive data on the device before it reaches an AI service.

Running your own IT? See ProctoredAI for your business