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

# Core concepts

> The vocabulary the rest of the docs use - harnesses, agent projects, kartas, sessions, sub-agents, the gateway, and releases.

Karta has a small, consistent vocabulary. Learn these eight terms and every
other page becomes easier to read.

If you are here to solve a specific job, use [Choose your path](/journeys) first;
come back here when a page uses a term you want to pin down.

<CardGroup cols={2}>
  <Card title="Agent harness" icon="robot">
    The runtime that executes the agent loop. See
    [Supported harnesses](/build/supported-harnesses) for the harness formats
    Karta detects and runs. The harness owns the model loop, tools, MCP
    connections, memory, context, and turns.
  </Card>

  <Card title="Agent project" icon="folder-tree">
    The deployable unit: a folder that configures a harness (`CLAUDE.md` or
    `AGENTS.md`, a native harness directory, sub-agents, skills, and
    `karta.toml`). The CLI and API use this term for the same object.
  </Card>

  <Card title="Karta" icon="fingerprint">
    A durable agent instance created from one agent project: one per user for a
    personal AI assistant, one per team virtual employee, one singleton for a
    backend task, or one per fleet member. Each karta has its own workspace and
    sessions.
  </Card>

  <Card title="Session" icon="comments">
    A lightweight orchestration handle - metadata, participants, current
    sub-agent, pending approvals. The session record is not the message
    store; conversation history lives with the harness.
  </Card>

  <Card title="Participant" icon="user-group">
    A human or AI actor in a session. Many can share one session; each message
    is attributed to its sender.
  </Card>

  <Card title="Sub-agent" icon="user-astronaut">
    An optional harness-native role inside one agent project, such as a
    `.claude/agents/*.md` file. It is packaged with the parent agent release,
    under the same agent endpoint.
  </Card>

  <Card title="Release" icon="box-archive">
    An immutable, versioned snapshot of an agent, materialized
    from a published artifact. Activation is a pointer flip.
  </Card>

  <Card title="Runtime and management boundary" icon="layer-group">
    Karta separates agent execution from account management. Running agent code
    does not become billing, key-management, or audit authority.
  </Card>
</CardGroup>

## How they fit together

```
end user / SDK
  |
  | kt_live_... key or session token
  v
KARTA RUNTIME
  | validates credential, checks budget
  | resolves session, karta, and active release
  v
HARNESS
  | runs the agent turn
  | streams typed events
  v
caller receives SSE or accumulated JSON

KARTA MANAGEMENT
  identity, API keys, usage, budgets, billing, BYOK, webhooks, audit
```

A request carries a credential, Karta validates it and checks the budget,
resolves (or creates) the session and karta, hands the turn to the harness, and
relays the harness's typed event stream back to the caller. Usage is metered
back to the account afterward. Walk through it in
[How Karta works](/concepts/how-karta-works#the-request-path).

## Who owns what

The harness owns the model loop, tools, MCP, memory, context, and conversation
history inside a running karta. Karta owns the platform around it: durable karta
identity, sessions, participants, multi-tenancy, releases, metering, and the
uniform API.

Because your agent stays in the harness's native format, Karta can read the
project layout directly.

## Read next

<CardGroup cols={2}>
  <Card title="Choose your path" icon="signs-post" href="/journeys">
    Route yourself by job before going deep on the concept model.
  </Card>

  <Card title="How Karta works" icon="sitemap" href="/concepts/how-karta-works">
    Per-session isolation, the request path, and the runtime/management boundary.
  </Card>

  <Card title="Agents" icon="folder-tree" href="/concepts/harness-applications">
    What you actually ship, and how a harness is detected.
  </Card>

  <Card title="Sessions & participants" icon="comments" href="/concepts/sessions-and-participants">
    Multi-turn state, multi-party conversations, and attribution.
  </Card>

  <Card title="Streaming events" icon="wave-pulse" href="/concepts/streaming-events">
    The typed event model every surface is built on.
  </Card>

  <Card title="Kartas & memory" icon="fingerprint" href="/concepts/instances-and-memory">
    Durable agent instances, memory, privacy, and how to choose the id that
    selects them.
  </Card>
</CardGroup>