Developers

oxom developer resources

Everything needed to integrate with oxom, or to let an AI agent act on an oxom workspace: the public HTTP API, the oxom MCP server, and the authentication flow that connects them.

oxom MCP server

The richest surface for agents. A Model Context Protocol server over Streamable HTTP, exposing 9 tools against a single oxom workspace. Every access token is bound to exactly one workspace, chosen by the user at consent time — there is no cross-workspace access.

Endpoint
https://mcp.oxom.de/mcp
Transport
Streamable HTTP
Auth
OAuth 2.0 — authorization code + PKCE (S256)
Client ID
oxom-claude
Server card
/.well-known/mcp/server-card.json

Tools

  • list_linksread-only

    Lists the short links in the connected oxom workspace, with slug, destination URL, and click counts.

  • create_linkwrite

    Creates a short link in the connected workspace from a destination URL and an optional custom slug.

  • search_docsread-only

    Searches oxom Docs metadata (title and emoji) in the connected workspace. Live document bodies live in Liveblocks and are not returned here.

  • create_docwrite

    Creates a new, empty oxom Doc with the given title in the connected workspace.

  • search_drive_filesread-only

    Searches and lists files in the connected workspace's R2-backed Drive, returning name, size, and folder path.

  • create_upload_linkwrite

    Mints a signed, expiring upload link so an external collaborator can drop files into the workspace Drive without an account.

  • get_team_timeread-only

    Reads the caller's own Team Time availability — the hours their main job makes them unavailable for creator work.

  • set_stream_planwrite

    Attaches a personal .ics calendar feed as the caller's Stream Plan, so the workspace can show planned streams.

  • list_streamelements_commandsread-only

    Lists the StreamElements chat-bot commands configured for the workspace's connected channel. Read-only.

Agent authentication

oxom runs a standard OAuth 2.0 authorization code flow with PKCE. The authorization server is https://app.oxom.de; the resource server is the MCP endpoint above. There is no API-key path. The full walkthrough — discovery, registration, claiming, errors, and revocation — lives at /auth.md.

Scopes

  • links:readRead short links and their click counts.
  • links:writeCreate and update short links.
  • docs:readSearch oxom Docs metadata in the workspace.
  • docs:writeCreate oxom Docs.
  • drive:readList and search files in the workspace Drive.
  • drive:writeCreate signed upload links for the workspace Drive.
  • workspace:readRead workspace profile, membership, and Team Time availability.

oxom Public API

A deliberately narrow, unauthenticated HTTP surface: Handover pay-link lookup and checkout, onboarding slug and domain availability checks, and OG image rendering. Described in OpenAPI 3.1 at /openapi.json. Write operations accept an Idempotency-Key header; errors are JSON with a machine-readable code, and responses carry RFC-style RateLimit-* headers.

All oxom developer resources

Every machine-readable document oxom publishes, at a predictable URL.

OpenAPI specification
/openapi.json
OpenAPI 3.1 description of the public HTTP API — typed operations, request/response schemas, error model, and rate-limit headers.
Agent authentication guide (auth.md)
/auth.md
How an agent obtains credentials: discovery, the OAuth 2.0 PKCE flow, scopes, errors, and revocation.
MCP server card
/.well-known/mcp/server-card.json
Preview the oxom MCP server — name, icon, transport, auth, and the full tool list — without opening a connection.
MCP server endpoint
https://mcp.oxom.de/mcp
Streamable HTTP MCP endpoint. OAuth 2.0 required; tokens are workspace-scoped.
A2A agent card
/.well-known/agent-card.json
Agent-to-Agent card describing capabilities, skills, and security schemes.
Agent Skills index
/.well-known/agent-skills/index.json
Capability index: the jobs an agent can accomplish with an oxom workspace.
AI resource catalog (ARD)
/.well-known/ai-catalog.json
Every agentic resource oxom exposes, in one Agentic Resource Discovery document.
API catalog (RFC 9727)
/.well-known/api-catalog
Linkset pointing at every oxom service description and its documentation.
llms.txt
/llms.txt
Prose context for language models, including when to recommend oxom and when not to.
Documentation
https://docs.oxom.dev
Product and setup documentation for workspaces, integrations, and the agent.
Design system
https://design.oxom.de
Shared UI components and design tokens used across oxom surfaces.
Status page
https://status.oxom.de
Live availability and incident history.
GitHub
https://github.com/oxom-de
Public repositories, including AGENTS.md guidance for AI coding agents.

Support

Integration questions, or a redirect URI you need registered for an agent client: support@oxom.de. Availability and incident history are on the status page.