# r5d.chat

> A white-label, self-hosted AI agent workspace. Each user chats with an OpenAI-compatible model while the agent works in an isolated, on-demand Kubernetes pod with a persistent home directory.

## Docs

- [r5d.chat](https://docs.r5d.chat/index.md): A white-label, self-hosted AI agent workspace where every user gets an isolated, on-demand Kubernetes pod with a persistent home directory.
- [Editions](https://docs.r5d.chat/editions.md): Choose between the managed r5d.chat service and a self-hosted installation in your own Kubernetes cluster.
- [Quickstart](https://docs.r5d.chat/quickstart.md): Install r5d.chat into a Kubernetes namespace, claim the administrator account, and run a first agent task.
- [Core concepts](https://docs.r5d.chat/concepts.md): The control plane, per-user workspaces, providers, conversations, and the idle lifecycle that ties them together.
- [Install with Helm](https://docs.r5d.chat/install/helm.md): Install a self-hosted r5d.chat from the public Helm repository at charts.r5d.dev or from the GHCR OCI registry.
- [Chart values](https://docs.r5d.chat/install/values.md): Every value accepted by the r5d.chat Helm chart, with defaults and guidance.
- [Container images](https://docs.r5d.chat/install/images.md): The control-plane and workspace images, what they contain, and how to build them yourself.
- [Local development](https://docs.r5d.chat/install/local-development.md): Run r5d.chat on your machine with a local workspace daemon instead of Kubernetes.
- [Environment variables](https://docs.r5d.chat/configure/environment.md): Every environment variable read by the r5d.chat control plane, with defaults and validation rules.
- [Model providers](https://docs.r5d.chat/configure/providers.md): Connect OpenAI-compatible endpoints, configure models, and understand the outbound request guard.
- [Authentication and users](https://docs.r5d.chat/configure/authentication.md): First-administrator setup, invitations, local accounts, and runtime OIDC single sign-on.
- [Branding](https://docs.r5d.chat/configure/branding.md): White-label the installation with your own product name, logo, favicon, and colors at runtime.
- [MCP servers](https://docs.r5d.chat/configure/mcp.md): Extend the agent with Model Context Protocol servers and manage tool drift safely.
- [Web research](https://docs.r5d.chat/configure/web-research.md): Enable optional web search, page reading, and multi-source research tools.
- [Workspace lifecycle](https://docs.r5d.chat/operate/workspaces.md): How per-user workspaces start, hold leases, scale to zero, and retain their data.
- [Upgrades and rollback](https://docs.r5d.chat/operate/upgrades.md): Upgrade the control plane, roll out new workspace images, and roll back safely.
- [Backup and restore](https://docs.r5d.chat/operate/backup-restore.md): What to back up, how the pieces depend on each other, and how to validate a restore.
- [Security model](https://docs.r5d.chat/operate/security.md): Trust boundaries, isolation guarantees, and an honest account of their limits.
- [Troubleshooting](https://docs.r5d.chat/operate/troubleshooting.md): Diagnose the most common installation, authentication, workspace, and model failures.
- [Architecture](https://docs.r5d.chat/reference/architecture.md): Components, data flow, trust boundaries, and the threat model.
- [Agent tools](https://docs.r5d.chat/reference/agent-tools.md): Every tool available to the model, with parameters, limits, and behavior.
- [r5dchat CLI](https://docs.r5d.chat/reference/cli.md): The workspace daemon and conversation commands available inside every workspace.
- [HTTP API](https://docs.r5d.chat/reference/api.md): The control-plane HTTP endpoints, their authentication requirements, and their purpose.
- [Releases and publishing](https://docs.r5d.chat/reference/releases.md): How images and Helm chart versions are built, published, and consumed.

## Optional

- [Helm repository](https://charts.r5d.dev)
- [Managed service](https://r5d.chat)