Guides
These guides cover choices the catalog cannot measure: ownership, memory fit, Windows, team servers, and data flow.
Choose
| Guide | What it answers | Updated | Page |
|---|---|---|---|
| Why own the agent runtime | One manifest, one root, a lock, checksums, and evidence: what oloop owns, what you own, and what removal actually deletes. | 2026-10-20 | Read the guide |
| Local coding model by memory size | What fits at 8, 16, 24, 32, 64, 128, and 192+ GB, which classes we measured, and where the catalog has no record. | 2026-10-20 | Read the guide |
Run
| Guide | What it answers | Updated | Page |
|---|---|---|---|
| Windows local AI: the stack and a coding agent | What Windows ML, Foundry Local, and Phi Silica do, what they do not give a coding agent, and the WSL2 path that works today. | 2026-10-20 | Read the guide |
Operate
| Guide | What it answers | Updated | Page |
|---|---|---|---|
| Self-hosted coding agent for a team | How a team runs one pinned coding setup on its own machines: shared manifests, one root per project, the measured L4 path, and a self-managed AWS L4. | 2026-10-20 | Read the guide |
Trust
| Guide | What it answers | Updated | Page |
|---|---|---|---|
| Secure local coding agent | What a local coding agent sends, what the Hub stores, the evidence you can keep, and the compliance claims we do not make. | 2026-10-20 | Read the guide |
Install
Install the CLI, then pick the setup for your machine class.
curl -fsSL https://oloop.run/install.sh | shQuestions
Why are the guides grouped this way?
The groups follow the decision a reader is making: whether to own the runtime, what fits, how to run it on a client or server, and what to tell a reviewer.
Are the guides a substitute for the machine pages?
No. A machine page carries a measured setup. A guide explains a choice or a boundary the catalog cannot prove yet, and links the measured pages where they exist.