Wowly Studios
MCP · V1

One endpoint for compatible AI clients.

The Wowly Studios MCP server exposes the same provider-neutral brand and publishing capabilities to ChatGPT, Codex, and other compatible hosts.

Production endpoint

https://wowlystudios.com/mcp

This is the canonical public HTTPS endpoint. Clients use the Model Context Protocol to initialize a session, discover tools, and call authorized capabilities.

Bring your own AI: the connected host performs reasoning and creative generation. The MCP path does not consume Wowly Studios AI credits or route requests through a Wowly-funded model.

Discovery

The server supports standard MCP initialization and tools/list. Tool names, descriptions, JSON input/output schemas, security schemes, and read/destructive/open-world annotations are generated from canonical V1 contracts.

The protected-resource discovery document is available at:

GET https://wowlystudios.com/mcp/.well-known/oauth-protected-resource

Authentication

Private data and actions require OAuth 2.1. The MCP server advertises its authorization server through protected-resource metadata. Compatible hosts complete an authorization-code flow with PKCE, then attach a bearer access token to tool calls.

  • The user signs in with their existing Wowly Studios Google account.
  • Tokens are audience-bound to the Wowly Studios MCP resource.
  • Every call resolves the connected user, connection, permissions, entitlement, and brand access.
  • Unauthenticated or out-of-scope requests fail closed.

Recommended workflow

  1. Call brands_list to obtain an accessible brand_id.
  2. Call brands_get_context before creating brand-specific content or claims.
  3. Use read tools to gather relevant assets, drafts, calendars, or performance.
  4. Save work as a private draft and revise it without destroying history.
  5. Preview scheduling with publications_schedule.
  6. Only after explicit user approval, call publications_confirm with the returned confirmation token.

Transport and formats

  • HTTPS remote MCP endpoint.
  • JSON-RPC MCP messages with structured JSON outputs.
  • Versioned tools generated from canonical JSON Schema contracts.
  • Idempotency keys on commands and optimistic concurrency tokens on mutable records.
  • Signed, short-lived media URLs rather than permanent public object URLs.

Next

Review the complete V1 tool catalog, read the security and confirmation model, or follow the ChatGPT connection guide.