# q15

> A self-hosted AI agent runtime for one person: persistent memory, a durable workspace,
> real command execution, and two clients, in a browser and in Telegram. Everything below is
> the documentation source as Markdown. `llms-full.txt` has all of it in one file.

## Start here

- [q15](https://q15.co/index.md): A self-hosted AI agent runtime for one person. Persistent memory, a durable workspace, real command execution, and two clients, in a browser and in Telegram.
- [Install q15](https://q15.co/install.md): Four commands on a server you control. The installer is in development; the published images are what runs today.
- [Why q15](https://q15.co/why-q15.md): What q15 claims, what each claim costs you, and what it does not claim at all.
- [Start here by intent](https://q15.co/start-here.md): Pick the line closest to what you want and read three pages instead of nineteen, instead of starting at the top of a table of contents.

## Using q15

- [Chat](https://q15.co/chat.md): Passkey sign-in, sealed chat, history and attachments, and the exact limits of what the browser client protects.
- [Telegram](https://q15.co/telegram.md): A second client for the same agent — a bot token, your numeric user ID, and one shared conversation.
- [Models](https://q15.co/models.md): How q15 chooses which model answers a turn, how to change it at runtime, and how to point it at a model you run yourself.
- [Memory](https://q15.co/memory.md): What q15 remembers, where each layer is stored, how it reaches the prompt, and how background jobs keep it current.
- [Workspace](https://q15.co/workspace.md): The durable project tree that survives restarts and upgrades.
- [Tools](https://q15.co/tools.md): The agent's capabilities — file operations, shell execution, web search, embeddings, media, skills, and subagents.
- [Skills](https://q15.co/skills.md): Reusable workflow packages that extend what the agent can do — markdown-defined, version-controlled, and portable across stacks.
- [Jobs](https://q15.co/jobs.md): Scheduled work that runs without you, with its own turn budget, its own tools, and a report back to the chat that created it.

## Under the hood

- [Architecture](https://q15.co/architecture.md): A turn end to end, the four services and why they are separate, and the storage contract that keeps the work.
- [Security](https://q15.co/security.md): What prompt injection is, why it is the fundamental challenge of LLM agents, and what q15's architecture does and does not defend against.
- [Credentials](https://q15.co/credentials.md): Two kinds of secret, two different homes, and the rules that keep the important ones out of the agent's reach.

## Operating it

- [Updates and rollbacks](https://q15.co/updating.md): How a q15 stack moves to another release, and how it goes back. The four images always move as one set.
- [Backups](https://q15.co/backups.md): What to copy, in what order, and the three things that bite you on restore.
- [Troubleshooting](https://q15.co/troubleshooting.md): Something is wrong with your q15 stack. Match the symptom, then read what to do about it.
- [Reference](https://q15.co/reference.md): Every key, path, port and command, in one place. Dry on purpose.