> ## Documentation Index
> Fetch the complete documentation index at: https://aidocs.ethanbragdon.icu/llms.txt
> Use this file to discover all available pages before exploring further.

# Starfire AI

> Learn the Starfire AI platform architecture, major product surfaces, feature status model, and where to begin.

# Starfire AI

Starfire AI is a **multi-model AI platform** for conversation, reasoning, research, knowledge, software creation, automation, and developer workloads.

It is designed around three primary ways of using the platform:

<CardGroup cols={3}>
  <Card title="Starfire AI" icon="sparkles" href="/chat/overview">
    The human-facing AI workspace for chat, research, projects, files, knowledge, and creation.
  </Card>

  <Card title="Control Center" icon="gauge-high" href="/admin/control-center">
    The operating surface for authorized administrators running the Starfire platform.
  </Card>

  <Card title="Developer Platform" icon="code" href="/developers/overview">
    APIs, credentials, applications, webhooks, logs, usage, and software integrations.
  </Card>
</CardGroup>

## One platform, multiple AI workloads

Starfire AI is more than a chat interface. The platform is being built as a shared system for several classes of work:

* **Chat & reasoning** — interactive conversations with selectable or routed AI models.
* **Projects** — persistent workspaces that organize conversations, files, knowledge, and outputs around a goal.
* **Knowledge** — files, repositories, retrieval, embeddings, and RAG-backed context.
* **Research** — search, source collection, claim verification, and citation-aware analysis.
* **FORGE** — code-oriented build workflows that can plan, create, validate, repair, and package software artifacts.
* **Agents & automation** — reusable AI roles plus task, workflow, schedule, watcher, approval, and memory foundations.
* **Organizations** — shared members, permissions, usage, billing, projects, and policies.
* **Developer workloads** — programmatic access to Starfire capabilities through a versioned developer platform.

## The Starfire architecture at a glance

```text theme={null}
                          ┌─────────────────────┐
                          │     Starfire AI     │
                          │  Product Platform   │
                          └──────────┬──────────┘
                                     │
          ┌──────────────────────────┼──────────────────────────┐
          │                          │                          │
          ▼                          ▼                          ▼
┌──────────────────┐      ┌──────────────────┐      ┌──────────────────┐
│ Human Workspace  │      │  Control Center  │      │ Developer        │
│                  │      │                  │      │ Platform         │
│ Chat             │      │ Users & Identity │      │ APIs             │
│ Projects         │      │ Organizations    │      │ API Keys         │
│ Knowledge        │      │ Models & Routing │      │ Apps             │
│ Research         │      │ Billing          │      │ Webhooks         │
│ FORGE            │      │ Observability    │      │ Logs & Usage     │
└────────┬─────────┘      └────────┬─────────┘      └────────┬─────────┘
         │                         │                         │
         └─────────────────────────┼─────────────────────────┘
                                   ▼
                    ┌──────────────────────────┐
                    │ Shared platform services │
                    │ identity · metering      │
                    │ model routing · storage  │
                    │ tools · policy · audit   │
                    └──────────────────────────┘
```

## Documentation status

Starfire AI is under active Alpha development. The documentation uses the following status language throughout the site:

* **Available** — implemented in the current product line and intended for normal use.
* **Alpha / staged** — implemented or actively shipping, but behavior, UI, or availability can still change.
* **Planned** — part of the documented product direction but not presented as generally available.

<Warning>
  A feature appearing in the documentation does not automatically mean every account can use it. Availability can depend on the active deployment, plan, organization policy, feature flags, model/provider health, and administrator rollout.
</Warning>

## Closed-source product

Starfire AI is a privately developed Starfire Technology Group product. These docs explain supported product behavior, administration, integrations, and public developer contracts. They intentionally do **not** publish private source code, infrastructure secrets, provider credentials, proprietary routing logic, or internal security implementation details.

## Start here

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Learn the fastest path from sign-in to a useful Starfire AI workflow.
  </Card>

  <Card title="Platform status & terminology" icon="circle-info" href="/getting-started/platform-status">
    Understand Alpha terminology, availability, and how to read these docs.
  </Card>

  <Card title="Projects & Knowledge" icon="folder-tree" href="/projects/overview">
    Learn how persistent workspaces, files, repositories, and retrieval fit together.
  </Card>

  <Card title="FORGE" icon="hammer" href="/forge/overview">
    Explore Starfire's software-building and artifact workflow.
  </Card>
</CardGroup>

## Go deeper

The documentation is organized so you can move from concepts into procedures and then into reference material.

<CardGroup cols={2}>
  <Card title="Choose a workflow" icon="signs-post" href="/getting-started/choose-workflow">
    Decide whether work belongs in Chat, Projects, Knowledge, Research, FORGE, Agents, or the Developer Platform.
  </Card>

  <Card title="Reference" icon="book-open" href="/reference/glossary">
    Use the canonical glossary, lifecycle states, limits, permission model, error taxonomy, and resource ID reference.
  </Card>

  <Card title="Operator runbooks" icon="gauge-high" href="/admin/user-investigation">
    Follow repeatable Control Center procedures for users, billing, models, providers, FORGE, Knowledge, security, and incidents.
  </Card>

  <Card title="Troubleshooting" icon="wrench" href="/troubleshooting/common-issues">
    Diagnose product problems by subsystem before changing permissions, billing, or configuration.
  </Card>
</CardGroup>