Agyn Documentation
Agyn is a Kubernetes-native platform for running AI agents. Users talk to agents in a chat interface; admins configure agents, models, tools, and secrets in the Console; operators run the platform on their own clusters. Every agent run is fully observable — every LLM call, tool execution, and context decision is recorded and inspectable.
These docs are organized around what you are trying to do.
Choose your path
| If you are… | Start with |
|---|---|
| A user talking to agents in chat | Use Agyn |
| An admin configuring agents, models, secrets, runners, apps for your organization | Administer |
| An operator installing or running Agyn on your own Kubernetes cluster | Production installation, then Operate |
| A developer integrating via API, Terraform, MCP, or building an app | Build & extend |
| New to Agyn | Introduction |
Sections
- Introduction — what Agyn is, core concepts, architecture at a glance.
- Production installation — install Agyn on your own Kubernetes cluster with the bootstrap Terraform stacks. Skip this section if you use Agyn Cloud.
- Administer — configure your organization, agents, models, secrets, runners, tools, and apps. Console UI and Terraform side by side.
- Use — everyday workflows: chat with agents, attach files, inspect runs, see usage, expose ports, manage API tokens.
- Build & extend — Gateway API, Terraform provider, MCP server authoring, agent CLI choice, apps development.
- Operate — day-2 operations: networking, identity, runners, scaling, backups, upgrades, security.
- Reference — glossary, service catalog, pointers to schemas and Helm values.
- Troubleshooting — diagnostic playbook by symptom and FAQ.
How these docs work
Each configuration topic follows the same pattern: a short concept block, Configure in the Console with screenshots and real route paths, Configure with Terraform with a working HCL snippet, and a related-links list. You can adopt either path — what the Console builds, Terraform can manage too.