generated: '2026-08-13' method: searched status: published source: https://docs.adverity.com/guides/mcp/mcp-setup.html deployment: mode: remote endpoint: https://mcp.eu.adverity.com/mcp auth: oauth verified: probed note: >- The remote endpoint is published verbatim in the Adverity MCP setup guide as the "Remote MCP server URL" to paste into Claude custom connectors and ChatGPT connectors. A POST of {"jsonrpc":"2.0","id":1,"method":"tools/list"} to it returned HTTP 401 invalid_token with a WWW-Authenticate: Bearer challenge naming https://mcp.eu.adverity.com/.well-known/oauth-protected-resource/mcp — so the server is live and the live tool schemas are auth-gated. No stdio/npx package is published; this is a hosted server only. probe: date: '2026-08-13' url: https://mcp.eu.adverity.com/mcp method: POST tools/list http_status: 401 body: invalid_token (OAuth bearer challenge) server: name: Adverity MCP product_name: Adverity Atlas MCP transport: http url: https://mcp.eu.adverity.com/mcp availability: beta access: >- Selected customers. Requires an active Adverity account with Management API (MAPI) access enabled; contact your Account Manager for the MCP beta program. clients: - Claude (Pro or higher, custom connectors) - ChatGPT (Business or Enterprise, developer-mode connectors) - Any MCP-compatible AI client authentication: model: oauth2-then-mapi-key note: >- The MCP client performs an OAuth 2.0 authorization-code flow (PKCE S256, dynamic client registration) against mcp.eu.adverity.com; at the Adverity sign-in page the user supplies their Instance URL and MAPI key. The AI assistant therefore inherits exactly the permissions of that Management API key — see scopes/adverity-scopes.yml. oauth_metadata: well-known/adverity-oauth-authorization-server.json protected_resource_metadata: well-known/adverity-oauth-protected-resource.json authorization_endpoint: https://mcp.eu.adverity.com/authorize token_endpoint: https://mcp.eu.adverity.com/token registration_endpoint: https://mcp.eu.adverity.com/register grant_types: [authorization_code, refresh_token] pkce: S256 safety: write_confirmation: two-phase note: >- Write operations follow a documented two-phase pattern: the tool first previews the change and applies it only after the user explicitly confirms. Destructive fan-out operations (replacing authorization account permissions, bulk-replacing datastream column mappings, deleting a destination or a user) are called out in the docs as full replacements that preview exactly what will be removed before applying. docs: - https://docs.adverity.com/guides/mcp/using-mcp.html - https://docs.adverity.com/guides/mcp/mcp-setup.html - https://www.adverity.com/updates/mcp-beta-program tools: - name: manage_datastream description: >- Lists, creates, and configures datastreams; view status, enable or disable pipelines, filter datastreams scheduled for deletion, explore connector types and their field options, and read or edit the metrics and dimensions a datastream collects. confirmation_required: [create, update, enable, disable, change-collected-fields] - name: manage_authorization description: >- Manages authorizations for connected data sources (Google Ads, Facebook Ads, LinkedIn Ads, etc.): create connections, check authorization status, handle OAuth flows, refresh metadata from external APIs, and manage which accounts an authorization may collect from including the grant-all flag. Has a list_permissions action. confirmation_required: [create, update-permissions, update-grant-all] - name: manage_fetch description: >- Controls data collection jobs — trigger fetches on demand or for a date range, monitor job status, stop running jobs, and review a datastream's fetch schedule. - name: manage_error description: >- Lists errors and their state, retrieves a single error by ID, retries failed operations, and acknowledges errors; retry/confirm can apply to one error or to every error in the same error group. confirmation_required: [retry, confirm] - name: manage_extract description: >- Lists and downloads data extracts from successful datastream executions; downloads return a preview of the first rows by default, full content on request. - name: manage_workspace description: Lists workspaces, shows their structure and hierarchy, lists available storage locations, and creates or updates workspaces. confirmation_required: [create, update] - name: manage_transformation description: >- Creates, lists, and edits transformations as part of datastream setup and retrieves the schema of supported processing operations; only name and description are editable after creation. confirmation_required: [create, update, delete] - name: manage_mapping_table description: >- Lists, retrieves, and updates mapping tables used for data harmonization and manages individual match-to-value entries; read-only access to value tables. Mapping tables cannot be created or deleted through the API. confirmation_required: [all-writes] - name: manage_data_mapping description: >- Maps a datastream's source columns to target columns — list source and target columns, create target columns, and map, unmap, or bulk-replace a datastream's column mappings. confirmation_required: [all-writes] - name: manage_destinations description: >- Creates and manages destinations where harmonized data is delivered (BigQuery, Snowflake, Google Sheets, etc.), lists destination types, and manages destination column mappings. confirmation_required: [create, update, delete] - name: manage_user description: Lists and manages instance users by ID — list, retrieve, create, update, delete — and sets any assignable platform role. confirmation_required: [create, update, delete] - name: manage_notifications description: >- Manages notification subscriptions across email, Slack, and Microsoft Teams — list, retrieve, create, update, delete — optionally scoped to specific datastreams or workspaces. confirmation_required: [create, update, delete] tool_count: 12 schemas: available: false note: >- Per-tool inputSchema could not be captured: tools/list requires an authenticated OAuth bearer token. Tool names and descriptions above are transcribed from the provider's published "Using Adverity MCP" guide, not invented. Live schema introspection needs an Adverity instance and MAPI key. related: - mcp/adverity-tool-crosswalk.yml - scopes/adverity-scopes.yml - authentication/adverity-authentication.yml