# Orca ## Docs - [Introduction](https://docs.orcapods.ai/introduction.md): Orca is a hosted platform for building, running, and publishing AI agents — with isolated workspaces, spend controls, and end-to-end observability built in. - [Quickstart](https://docs.orcapods.ai/quickstart.md): Create your account, build your first agent in the dashboard, run it, then call it from the API. - [Agent Profiles](https://docs.orcapods.ai/concepts/profiles.md): Reusable agent templates that define how an agent thinks, which tools it can use, and which LLM it runs on. - [Sessions](https://docs.orcapods.ai/concepts/sessions.md): Live instances of an agent profile, scoped to a single runner and equipped with their own MCP endpoint. - [Runs](https://docs.orcapods.ai/concepts/runs.md): A Run is a single task execution against a session. Events stream in real time over SSE. - [Agent Pools](https://docs.orcapods.ai/concepts/pools.md): Named groups of profiles that share a filesystem workspace and a lightweight blackboard. - [Workflow Orchestration](https://docs.orcapods.ai/concepts/workflows.md): DAG-based multi-agent workflows executed by the conductor-side planner engine. - [Skills](https://docs.orcapods.ai/concepts/skills.md): Reusable Agent Skills — packaged instructions plus scripts, references, and assets that an agent loads and runs in its sandbox. - [Memory Bank](https://docs.orcapods.ai/concepts/memory-bank.md): Per-profile long-lived memory: structured saves, multi-signal relevance ranking, and automatic prompt injection. - [MCP Integration](https://docs.orcapods.ai/concepts/mcp-integration.md): How Orca uses the Model Context Protocol to give agents a unified interface for platform tools and external APIs. - [Connected Apps](https://docs.orcapods.ai/concepts/connected-apps.md): OAuth into third-party SaaS apps via Composio and expose them to agents as managed MCP servers — no manual MCP configuration. - [Publishing](https://docs.orcapods.ai/concepts/publishing.md): Public chat-gateway routes for tenant-owned agent profiles: data model, identity, conversations, and the gateway-to-conductor split. - [Storage and Files](https://docs.orcapods.ai/concepts/storage.md): VirtualFS-backed filesystem tools, shared storage mounts, and run artifact handling. - [Sandboxes](https://docs.orcapods.ai/concepts/sandboxes.md): Optional per-agent compute environments allocated when a session starts. Surfaced in the dashboard as Environments. - [Secrets](https://docs.orcapods.ai/concepts/secrets.md): A per-tenant encrypted secret vault. Reference values as secret:// in MCP headers and file mounts — plaintext is never displayed after creation. - [API Keys](https://docs.orcapods.ai/concepts/api-keys.md): Durable ao_ tenant keys for authenticating SDK, API, and automation calls — scoped to an environment and inheriting their creator's role. - [Credits & Billing](https://docs.orcapods.ai/concepts/billing.md): Prepaid credits, spend caps that gate runs, and per-run usage metering. - [Roles & Access](https://docs.orcapods.ai/concepts/access-control.md): Organizations isolate tenant data; role-based access control (owner, admin, member, viewer) governs who can do what. - [Local Development](https://docs.orcapods.ai/guides/local-development.md): Set up a full Orca development environment with hot reload on all components. - [Docker Deployment](https://docs.orcapods.ai/guides/docker-deployment.md): Run the Orca stack with Docker Compose for production-parity local environments. - [Kubernetes Deployment](https://docs.orcapods.ai/guides/kubernetes.md): Scale Orca to production with Kubernetes StatefulSets for runners and Deployments for conductors. - [Multi-Runner Setup](https://docs.orcapods.ai/guides/multi-runner.md): Distribute agent sessions across multiple runners for scalability and runtime specialization. - [Publishing an Agent](https://docs.orcapods.ai/guides/publishing.md): End-to-end walkthrough: configure the chat-gateway, publish a profile, mint an API key, hit the public endpoint, and operate the surface. - [Observability](https://docs.orcapods.ai/guides/observability.md): Prometheus metrics, structured logs, and OpenTelemetry traces for Orca services. - [Use Orca from Claude Code](https://docs.orcapods.ai/agents/claude-code.md): Install the Orca plugin, sign in headlessly, and drive cloud agents from a Claude Code session. - [Use Orca from Cursor](https://docs.orcapods.ai/agents/cursor.md): Register the Orca MCP server in Cursor and drive cloud agents from the editor. - [Use Orca from Codex](https://docs.orcapods.ai/agents/codex.md): Register the Orca MCP server in Codex CLI and drive cloud agents from it. - [The Orca MCP server](https://docs.orcapods.ai/agents/mcp.md): orca mcp serve exposes the whole control plane as MCP tools: reference for every tool, auth resolution, and the api_request escape hatch. - [Headless authentication](https://docs.orcapods.ai/agents/headless-auth.md): Sign in to Orca from terminals with no browser: the device flow, environment variables, and CI patterns. - [CLI for agents](https://docs.orcapods.ai/agents/cli.md): The orca CLI's machine contract: JSON everywhere, NDJSON streams, stable exit codes, env-only auth. - [Dashboard overview](https://docs.orcapods.ai/dashboard/overview.md): Sign in, pick an active organization, navigate the sidebar, and use the command palette, topbar, and role-gated surfaces. - [Workbench, Chat & Compare](https://docs.orcapods.ai/dashboard/workbench.md): Run an agent against a prompt and inspect every streamed step — in Workbench, Chat, or Compare mode. - [Agents](https://docs.orcapods.ai/dashboard/agents.md): Create, import, edit, publish, and manage the full lifecycle of your agent profiles. - [Pools](https://docs.orcapods.ai/dashboard/pools.md): Group agents into a shared workspace with role-aware access-control lists. - [Skills](https://docs.orcapods.ai/dashboard/skills.md): Import, preview, and manage Agent Skills that run in the sandbox. - [Memory Bank](https://docs.orcapods.ai/dashboard/memory-bank.md): Browse, search, and manage the memories agents write at runtime. - [Sessions](https://docs.orcapods.ai/dashboard/sessions.md): Inspect live and past sessions across every agent, with their runs, streams, and timelines. - [Workflows](https://docs.orcapods.ai/dashboard/workflows.md): Inspect, execute, and cancel multi-agent DAG workflows. - [MCP Servers](https://docs.orcapods.ai/dashboard/mcp-servers.md): Register MCP servers, connect SaaS apps via OAuth, and manage active connections. - [Files](https://docs.orcapods.ai/dashboard/files.md): Browse the virtual filesystem and manage storage mounts. - [Usage](https://docs.orcapods.ai/dashboard/usage.md): Credit balance, spend cap, and windowed activity for your organization. - [Stats](https://docs.orcapods.ai/dashboard/stats.md): Auto-refreshing fleet-level metrics for agents and sandbox environments. - [Runtime](https://docs.orcapods.ai/dashboard/runtime.md): Inspect the sandbox workers leased to an organization. - [Workspace & members](https://docs.orcapods.ai/dashboard/workspace.md): Organization settings, member management, and your personal account. - [Environments](https://docs.orcapods.ai/dashboard/environments.md): Sandbox compute leases attached to sessions, with an audit timeline. - [Storage](https://docs.orcapods.ai/dashboard/storage.md): The shared S3-compatible artefact bucket where agents publish outputs. - [Secrets](https://docs.orcapods.ai/dashboard/secrets.md): Admin-only tenant secret vault, referenced elsewhere as secret://. - [API Keys](https://docs.orcapods.ai/dashboard/api-keys.md): Issue, reveal, and revoke durable ao_ tenant API keys for SDKs, CI, and scripts. - [Audit](https://docs.orcapods.ai/dashboard/audit.md): Review agent lifecycle and configuration changes with before/after diffs. - [Quickstart templates](https://docs.orcapods.ai/dashboard/quickstart.md): Create your first agent or pool from a pre-filled template. - [SDKs](https://docs.orcapods.ai/sdk/overview.md): Use the generated Orca SDKs to create profiles, launch runs, stream events, and manage platform resources from application code. - [TypeScript SDK](https://docs.orcapods.ai/sdk/typescript.md): Create profiles, submit runs, stream events, and manage Orca from TypeScript. - [Python SDK](https://docs.orcapods.ai/sdk/python.md): Use the sync and async Python clients to manage Orca and consume run streams. - [Go SDK](https://docs.orcapods.ai/sdk/go.md): Status of the Go client, and how to call the Orca API from Go today. - [SDK Method Map](https://docs.orcapods.ai/sdk/reference.md): Cross-language map of generated TypeScript, Python, and Go SDK resource methods. - [VirtualFS](https://docs.orcapods.ai/virtualfs/overview.md): Tenant-scoped, sealed virtual filesystem shared by agents, the conductor, the dashboard, and platform services. - [VirtualFS SDKs](https://docs.orcapods.ai/virtualfs/sdks.md): Use the standalone Python and TypeScript clients, the Claude Bash adapter, and the OpenAI Agents sandbox adapter to call the VirtualFS HTTP API. - [VirtualFS API](https://docs.orcapods.ai/virtualfs/api.md): HTTP API exposed by the standalone virtual filesystem server for typed file operations, mount discovery, tenant KB files, cached directory listings, and sealed command execution. - [Conductor API](https://docs.orcapods.ai/api-reference/conductor.md): HTTP API exposed by the Orca Conductor, the stateless control plane used by the dashboard and external clients. - [Chat Gateway API](https://docs.orcapods.ai/api-reference/chat-gateway.md): Public HTTP API for tenant-published agents. - [Runner Wire API](https://docs.orcapods.ai/api-reference/runner.md): Internal HTTP API exposed by each Runner for conductor routing, session ownership, tool invocation, and MCP access. - [Agent Worker API](https://docs.orcapods.ai/api-reference/sidecar.md): HTTP API for the agent-worker Node.js sidecar that executes LLM inference. - [Platform Tools Overview](https://docs.orcapods.ai/platform-tools/overview.md): Built-in tools hosted by every runner and exposed through each session's scoped MCP endpoint. - [Introspection Tools](https://docs.orcapods.ai/platform-tools/introspection.md): Platform tools for runner/session metadata, profile discovery, usage, topology, and web lookup. - [Orchestration Tools](https://docs.orcapods.ai/platform-tools/orchestration.md): Profile-scoped tools for delegation, run monitoring, and DAG plan control. - [Architecture](https://docs.orcapods.ai/concepts/architecture.md): The two-tier conductor/runner design that makes Orca horizontally scalable.