| Guides | Get started | Introduction | index.mdx | Deploy an agent defined for a supported harness as durable kartas with sessions, releases, identity, budgets, webhooks, and audit controls. |
| Guides | Get started | Choose your path | journeys.mdx | Find the right route through the Karta docs for your current job: first deploy, existing harness project, embedded widget, backend API, operations, or enterprise review. |
| Guides | Get started | How-to index | how-to.mdx | Search-friendly answers for the public Karta jobs in the process inventory, with each question routed to the canonical docs page. |
| Guides | Get started | Table of contents | table-of-contents.mdx | A complete route and folder map for every public Karta docs page. |
| Guides | Get started | Quickstart | quickstart.mdx | Deploy a starter agent, or an existing agent defined for a supported harness, to Karta in under 10 minutes. |
| Guides | Get started | Overview | concepts/overview.mdx | The vocabulary the rest of the docs use - harnesses, agent projects, kartas, sessions, sub-agents, the gateway, and releases. |
| Guides | Get started | Glossary | glossary.mdx | Definitions for the core Karta terms used across the docs. |
| Guides | Concepts | How Karta works | concepts/how-karta-works.mdx | What happens when a request reaches your agent on Karta - durable kartas, per-session isolation, the request path, and the separation between running agents and managing them. |
| Guides | Concepts | Agents | concepts/harness-applications.mdx | What you actually ship: an agent project that a harness can run - and how Karta detects, packages, and runs it. |
| Guides | Concepts | Sessions & participants | concepts/sessions-and-participants.mdx | Multi-turn conversation handles, multi-party messaging, and per-sender attribution - without making the session record a message store. |
| Guides | Concepts | Sub-agents | concepts/agents.mdx | Harness-native sub-agents defined inside one agent project, discovered automatically, routed and handed off at runtime. |
| Guides | Concepts | Kartas & memory | concepts/instances-and-memory.mdx | The id you choose selects the durable agent instance, workspace, and memory a session uses - one per user, virtual employee, backend task, or fleet member. |
| Guides | Concepts | Streaming events | concepts/streaming-events.mdx | How Karta streams turns across native sessions, the Managed Agents adapter, and the OpenAI-compatible Responses adapter. |
| Guides | Concepts | Multi-tenancy | concepts/multi-tenancy.mdx | How Karta isolates one organization’s data from another’s - auth-enforced org scoping, durable karta boundaries, and per-session isolated runtime environments. |
| Guides | Concepts | Releases | concepts/releases.mdx | Immutable, versioned snapshots of an agent - built once, activated by an atomic pointer flip, rolled back the same way. |
| Guides | Agent project | Supported harnesses | build/supported-harnesses.mdx | The harness formats Karta can detect, scaffold, package, and run. |
| Guides | Agent project | Agent structure | build/agent-structure.mdx | Lay out an agent file by file - the one required file, plus optional sub-agents, skills, and Karta config. |
| Guides | Agent project | Sub-agent files | build/defining-agents.mdx | The Markdown-with-frontmatter format for harness sub-agents across supported harnesses. |
| Guides | Agent project | Skills | build/skills.mdx | Reusable agent capabilities, defined in your harness’s native skill format and discovered automatically. |
| Guides | Agent project | karta.toml | build/karta-toml.mdx | An agent’s local identity, deploy gate, and build hints. |
| Guides | Agent project | Configuration | build/configuration.mdx | karta.jsonc - optional CLI, harness, and runtime settings that tune how your agent behaves at run time. |
| Guides | The CLI | CLI overview | cli/overview.mdx | Karta CLI commands for setup, local parity, deploy, rollback, logs, and account operations. |
| Guides | The CLI | Command reference | cli/reference.mdx | The full top-level karta CLI surface: setup, deploy, observe, account, API, and management commands. |
| Guides | The CLI | Manage agents | cli/agents-and-sessions.mdx | Deploy agents, inspect releases, roll back, and stream logs - karta.toml links your local project to its hosted agent. |
| Guides | The CLI | Workspace access | cli/workspace-access.mdx | Clone a karta’s durable files read-only over git, once an admin enables workspace access for the agent. |
| Guides | Deploy | The deploy loop | deploy/deploy-loop.mdx | Ship, serve, consume: publish an agent, get a stable agent path, and route sessions into durable kartas. |
| Guides | Deploy | Releases & rollback | deploy/releases-and-rollback.mdx | Deploy a new release and roll back to a previous one - with a git push, from CI, and over the API. |
| Guides | Platform | Production readiness | platform/production-readiness.mdx | A launch checklist for deploying Karta agents with the right access, spend, release, observability, webhook, and support controls. |
| Guides | Platform | API keys | platform/api-keys.mdx | Server-side kt_live_ bearer tokens, least-privilege scopes, expiration rules, and safe rotation. |
| Guides | Platform | Usage, cost & budgets | platform/usage-and-budgets.mdx | How Karta meters tokens and cost, the reports and alerts you get, and the hard caps that return 402 before you run over budget. |
| Guides | Platform | Billing & credits | platform/billing-and-credits.mdx | Prepaid credits, how usage draws them down, and Karta’s guarantee that you can never owe more than you’ve paid. |
| Guides | Platform | BYOK | platform/byok.mdx | Use your own model-provider key or compatible endpoint, select a model and portable parameters, and rotate the connector when needed. |
| Guides | Platform | Webhooks | platform/webhooks.mdx | Receive signed, retried event notifications from Karta, verify signatures, and recover failed endpoints. |
| Guides | Platform | Audit log | platform/audit-log.mdx | An immutable record of material account changes, exportable through the dashboard and API. |
| Guides | Platform | Troubleshooting & FAQ | platform/troubleshooting.mdx | Common problems running agents on Karta, what causes them, and how to fix them - plus where to check status and how to reach us. |
| Guides | Security | Production architecture review | security/architecture-review.mdx | A review path for platform, security, and enterprise teams evaluating how Karta runs agents in production. |
| Guides | Security | Security model | security/security-model.mdx | How Karta thinks about prompt injection, execution isolation, credentials, tools, egress, operator access, and residual risk. |
| Guides | Security | Trust & compliance | security/trust.mdx | Encryption, authentication hardening, isolation, and the platform’s security posture. |
| Guides | Security | Session tokens | security/session-tokens.mdx | The short-lived, agent-scoped credential your backend mints so an end user’s browser can talk to a Karta agent - without ever holding an API key. |
| Guides | Security | Chat widget | security/chat-widget.mdx | The CSP directives your page needs, the origin-allowlist and abuse model, and the data path for the embeddable chat widget. |
| Guides | Enterprise | BYOC | enterprise/byoc.mdx | Run Karta’s data plane inside your own AWS account. Workspaces, durable state, and model traffic stay in your cloud; Karta operates the control plane. |
| Guides | Tutorials | Run a support bot locally | tutorials/support-bot.mdx | A complete local Karta-session walkthrough - three personas, a streaming chat widget, no cloud. Operator runs Karta, an agent builder brings a support bot project, an end user chats. |
| SDKs | Python SDK | Overview | sdks/python/overview.mdx | karta-runtime - embed Karta in-process or serve it over HTTP. |
| SDKs | Python SDK | Karta | sdks/python/karta.mdx | The orchestrator - construct it, discover agents, send and stream turns, register hooks, set policies. |
| SDKs | Python SDK | Session | sdks/python/sessions.mdx | The multi-turn handle returned by app.session() - participants, agent handoff, send/stream, and resolving approvals. |
| SDKs | Python SDK | Hooks & policies | sdks/python/hooks-and-policies.mdx | React to lifecycle events, and gate messages against configurable policies. |
| SDKs | TypeScript | Management SDK | sdks/typescript/management.mdx | The TypeScript client for account operations - API keys, BYOK, webhooks, usage, and billing. |
| SDKs | TypeScript | CLI | sdks/typescript/cli.mdx | The karta CLI on npm - browser sign-in, setup, local parity, git-push deploys, rollback, and logs. |
| SDKs | Chat Widget | Quickstart | sdks/widget/quickstart.mdx | Drop a hosted, themeable agent chat widget onto any site with one async script tag. |
| SDKs | Chat Widget | Command API | sdks/widget/command-api.mdx | Drive the widget from your page with karta(…) - open it from your own button, identify the user, send messages, and subscribe to events. |
| SDKs | Chat Widget | Identity | sdks/widget/identity.mdx | Three levels of end-user identity in the widget - anonymous, soft, and verified (HMAC) - and exactly how the verified scheme works. |
| SDKs | Chat Widget | Authenticated agent | sdks/widget/authenticated-agent.mdx | Embed a Karta agent for signed-in users with verified identity, trusted per-user context, and host-attested step-up for sensitive approvals. |
| SDKs | Chat Widget | Appearance & config | sdks/widget/theming.mdx | Style and configure the widget from the agent’s Appearance and Embed tabs - colors, type, layout, icons, display modes, transcript detail, approvals, origins, and pre-chat consent. |
| SDKs | Chat Widget | Headless & React | sdks/widget/headless-react.mdx | Skip the prebuilt UI: build your own chat on the headless KartaAgentClient, or use the @karta/react bindings. |
| API Reference | Overview | Introduction | api-reference/introduction.mdx | The Karta HTTP API - sessions, streaming, gateway, and OpenAI- and Anthropic-shaped consumer adapters. |
| API Reference | Overview | Authentication | api-reference/authentication.mdx | Two credential types - API keys and session tokens - and where each one is valid. |
| API Reference | Overview | Errors | api-reference/errors.mdx | HTTP status codes Karta returns, what each means, and how errors surface mid-stream. |
| API Reference | Sessions | Sessions | api-reference/sessions.mdx | Create, fetch, and resume sessions - on the flat route family and the agent route family. |
| API Reference | Sessions | Messages | api-reference/messages.mdx | Send a turn into a session - accumulated JSON, or a streamed sequence of typed SSE events. |
| API Reference | Sessions | Inputs | api-reference/inputs.mdx | Resolve an approval prompt a turn paused on - approve once, approve for the session, or deny. |
| API Reference | Sessions | Gateway | api-reference/gateway.mdx | Submit events into the gateway and deliver them to participants - the routing contract for multi-participant messaging, including participants reachable at another gateway. |
| API Reference | Consumer adapters | Overview | api-reference/adapters/overview.mdx | Talk to a Karta agent with OpenAI- or Anthropic-shaped client code - same turn, re-framed into familiar wire formats. |
| API Reference | Consumer adapters | OpenAI Responses | api-reference/adapters/openai-responses.mdx | Drive a Karta agent with the OpenAI Responses API shape, including previous_response_id continuation. |
| API Reference | Consumer adapters | OpenAI Chat Completions | api-reference/adapters/openai-chat-completions.mdx | Drive a Karta agent with the classic OpenAI Chat Completions messages-array shape. |
| API Reference | Consumer adapters | Anthropic Managed Agents | api-reference/adapters/managed-agents.mdx | Event-sourced managed-agent sessions - post user events, stream a replayable agent.* event log, and confirm tool use. |