AI Agents (MCP Connector)

Last updated: 2026-10-23Connect your APERTURESyndicate account to Claude, ChatGPT or another AI agent — what it can see and do, what it never gets, and how to cut it off.

AI Agents (MCP Connector)

Effective date: 23 October 2026 · Version 1.0

You can connect your APERTURESyndicate account to an AI agent — Claude, ChatGPT, Claude Code or any other client that speaks the Model Context Protocol (MCP). The agent can then answer questions about your account and act in it when you ask: build a Resonance playlist, update your profile, post on Shitpost, create an ASAI character.

These terms are product-specific. Under Section 3 of the Terms of Service they sit below the platform-wide documents; where they appear to conflict with the Terms of Service, the Acceptable Use Policy or the Privacy Policy, the platform-level document controls.

1. How to connect

  • Server address: https://mcp.aperturesyndicate.com/mcp
  • Claude (claude.ai, desktop, mobile): Settings → Connectors → find APERTURESyndicate in the directory, or add a custom connector with the address above.
  • Claude Code: claude mcp add --transport http aperturesyndicate https://mcp.aperturesyndicate.com/mcp, then /mcp to sign in.
  • Other MCP clients: add the address as a remote (Streamable HTTP) server.

The client sends you to our sign-in page. After you sign in, a consent screen shows the agent's name, the address it will return you to, and every permission it asks for. Nothing is shared until you press Allow.

2. What the agent can do

Each permission on the consent screen unlocks a set of tools. The agent only sees tools for the permissions you allowed.

PermissionWhat the agent can do
See your accountRead your profile, badges, subscriptions, stars and quests, Resonance listening stats, Rebellion and Apeiron playtime, communities, notifications, your Shitpost posts, creator wallet and ASAI energy
Change your profileChange nickname, bio, links, avatar and banner
ApplicationsSend developer and creator applications in your name
ASAICreate characters in your ASAI library (private until you publish them)
ApeironSet up store pages, builds and releases of your games and mods, and send them for review
ResonanceSearch the catalogue, create and edit your playlists
ShitpostPublish posts as you and delete your own posts

The agent acts as you and goes through the same checks as the website: it cannot reach other people's private data, cannot post as the Syndicate or as a community, and cannot accept agreements for you — application tools require you to confirm consent yourself. Changes are limited per hour.

3. What the agent never gets

  • Your password, two-factor secrets or session cookies.
  • Your email address and internal username — the agent's token does not carry them.
  • Access to any other APERTURESyndicate API. Its token is accepted only by the MCP server.

4. Your data and the agent's provider

Whatever a tool returns is sent to the AI agent you connected, and from there to its provider (for example Anthropic for Claude, OpenAI for ChatGPT), which processes it under its own terms and privacy policy. We send it only when the agent calls a tool on your behalf. The provider is not our sub-processor: you choose it and you direct it, as described in the Privacy Policy.

Write tools change your account exactly as the website would. Posts the agent publishes are public. Review what the agent proposes before you let it act — AI agents can make mistakes (see AI Output & Limitations).

5. Disconnecting

Open Settings → Security on APERTURESyndicate. Each connected agent is listed as "<name> · AI agent" among the places you are signed in. Sign out cuts it off within a minute. Signing out everywhere or changing your password disconnects every agent too.

6. Support

Questions or problems: Support or [email protected].

AI Agents (MCP Connector) | AS Docs