name: Arcade Vocabulary description: > Controlled vocabulary covering the operational and capability dimensions of the Arcade.dev developer platform — the Arcade Engine API, the MCP Server framework, the agent-side SDKs, and the auth-provider catalog. Terms are sourced from the public OpenAPI spec at https://api.arcade.dev/v1/swagger, the docs at https://docs.arcade.dev, and the ArcadeAI GitHub org. url: https://github.com/api-evangelist/arcade created: '2026-05-22' modified: '2026-05-22' domains: - name: Platform description: Top-level concepts naming Arcade as a category and product. terms: - term: MCP Runtime description: > Arcade's category framing — "the only MCP runtime for secure, reliable AI agent deployments." A runtime hosts Model Context Protocol servers, mediates user authorization, and routes tool calls from agent frameworks to downstream APIs. - term: Arcade Engine description: > The HTTP API surface (api.arcade.dev) that exposes tool catalogs, authorization flows, execution endpoints, hooks, gateways, plugins, and worker management. Documented by the OpenAPI 3.0 spec at api.arcade.dev/v1/swagger. - term: Arcade Registry description: > The marketplace (in beta) where publishers list and monetize MCP-compatible tools and toolkits consumable by agent runtimes. - term: Arcade Cloud description: > Arcade's managed deployment target — hosted MCP servers, secrets storage, dashboards, and metered execution covered under the Hobby, Growth, and Enterprise plans. - name: Tool Model description: Concepts in the agent-tool execution model. terms: - term: Tool description: > A single invocable capability addressed by a fully-qualified name and backed by an input/output schema, an authorization requirement, and an execution worker. - term: Toolkit description: > A versioned bundle of related tools (e.g. Gmail, Linear, PostHog) packaged for installation into an Arcade-hosted or self-hosted MCP server. - term: Tool Definition description: > The structural contract for a tool — name, description, input/output parameters, authorization requirement, and toolkit metadata — published in the ArcadeAI/schemas repo. - term: Tool Execution description: > A single run of a tool — request payload, authorization context, worker attempt history, and the formatted response returned to the agent. - term: Scheduled Tool description: > A tool execution registered to run on a schedule rather than synchronously in response to an agent call. - term: Pro Tool Execution description: > Plan-metered execution class for higher-cost or higher-effort tool calls; billed separately from standard executions on the Growth tier. - term: Formatted Tool description: > A tool rendered for a target agent-framework consumer (LangChain, OpenAI Agents, CrewAI, Vercel AI, Mastra, etc.). - name: Authorization description: User-, app-, and agent-level authorization concepts. terms: - term: Auth Provider description: > An OAuth or generic identity integration registered on Arcade — Asana, Atlassian, Discord, Dropbox, GitHub, Google, HubSpot, Linear, LinkedIn, Microsoft, Notion, Reddit, Slack, Spotify, Twitch, X, Zoom, Airtable, Attio, Calendly, ClickUp, Figma, Mailchimp, Miro, PagerDuty, Salesforce, Square, TickTick, Zendesk, Zoho, plus the generic OAuth 2.0 provider. - term: Authorization Requirement description: > The set of scopes, providers, and binding type a tool needs before it can execute on behalf of a specific user. - term: Authorization Context description: > The resolved token, identity, and scope bundle that accompanies a single tool execution. - term: User Connection description: > A persisted link between an Arcade user and an upstream identity (Google, Slack, etc.) plus the tokens that authorize tool calls on their behalf. - term: Binding Type description: > How a credential is bound to an execution — per-user, per-tenant, or shared service credential. - term: User Challenge description: > A unit of metered interactive consent — the plan-quoted cost when Arcade asks an end user to grant or refresh an authorization. - term: Custom User Verifier description: > A customer-supplied verifier (e.g. Stytch, Supabase) that Arcade calls before activating a token for a tool call. - term: Session Verification description: > Admin-level policy controlling how Arcade re-verifies an active session before sensitive tool calls. - name: Runtime And Topology description: Deployable runtime components and their topology. terms: - term: Worker description: > A process that hosts and executes tools. Workers can be Arcade-hosted or self-hosted; the Engine routes tool calls to a worker, optionally through a gateway. - term: Gateway description: > An ingress in front of one or more workers within a project, applying routing rules, headers, and policies. - term: Hook description: > A configurable extension point — pre- or post-call — that injects custom logic into the tool execution pipeline. - term: Hook Point description: > The position in the execution lifecycle at which a hook runs. - term: Plugin description: > A registered extension (typically a hook bundle or auth verifier) that plugs into the Engine for an org. - term: MCP Server description: > A Model-Context-Protocol-compatible server exposing tools; can be Arcade-hosted (counted against the plan's hosted-server quota) or self-hosted (via arcade-mcp). - term: Project description: > A logical container inside an organization that owns a set of gateways, workers, and tools. - term: Organization description: > The Arcade tenant — owns projects, billing, and identity policies. - name: Observability description: Visibility and quality concepts. terms: - term: Eval description: > A scripted evaluation of tool-calling behavior, runnable through the `arcade evals` CLI command and used to gate releases. - term: Execution Log description: > Recorded request, response, and authorization context for a tool execution — the basis of Arcade's observability surface. - term: Audit Log description: > Enterprise-tier event log used for compliance and incident review. - term: Status Page description: > status.arcade.dev — the public uptime surface; reports "All Systems Operational" with 100% uptime over the past 90 days as of 2026-05-22. - name: Commercial description: Plans and pricing concepts. terms: - term: Hobby description: > Free tier — 100 user challenges, 1,000 standard tool executions, 50 pro tool executions, one Arcade-hosted MCP server, community GitHub support. - term: Growth description: > $25/month tier — 600 user challenges (then $0.05 each), 2,000 standard executions (then $0.01 each), 100 pro executions (then $0.50 each), hosted MCP servers at $0.05/hour, email support with SLA. - term: Enterprise description: > Custom-priced tier — dedicated tenant isolation, audit logs, RBAC, SSO/SAML, dedicated account rep, custom SLAs. - term: Startup Program description: > Custom pricing track for companies under 100 employees, nonprofits, and educational institutions; entry via contact@arcade.dev.