# Agento Desktop documentation

## For users

| Guide | What it covers |
| --- | --- |
| [Installation](/docs/installation/) | Downloads per platform, first launch, updates, where your data lives, uninstalling |
| [User Guide](/docs/user-guide/) | Every section of the app: chats, agents, integrations, scheduled tasks, session history, analytics, the LLM gateway, settings |
| [Troubleshooting](/docs/troubleshooting/) | Common problems, and how to read the logs |

Start with [Installation](/docs/installation/).

## For contributors

| Guide | What it covers |
| --- | --- |
| [Architecture](/docs/architecture/) | Stack, process model, the native backend, the Claude SDK, design principles, and the frozen wire-format spec |
| [Development](/docs/development/) | Setup, running locally, tests, parity testing, conventions, debugging |
| [Releasing](/docs/releasing/) | Cutting a release, release notes, the guards, the update manifest, signing keys |

[`CLAUDE.md`](https://github.com/shaharia-lab/agento/blob/main/CLAUDE.md) is the working notes, and it is the repository's
only one — the root file holds the cross-cutting rules and indexes the
per-subsystem notes in `docs/internal/`: every decision, with the reasoning and
the failures behind it. These guides are the map, that file is the territory.

`docs/internal/` is those subsystem notes, written for maintainers and for AI
agents working on the code, one file per area. It is deliberately not published
to the documentation site — the guides above are.