Add the Fanduit MCP server to Claude and it gets 363 tools across every connected creator. Ask in plain language, and Claude reads live data, drafts replies, and pulls reports.
# point Claude at the Fanduit MCP server { "mcpServers": { "fanduit": { "command": "php", "args": ["artisan", "mcp:serve"], "env": { "FANDUIT_API_KEY": "sk_live_…" } } } }
The whole surface, as tools.
Claude calls real tools, so answers reflect the account right now, not a stale export.
Have Claude read a chat, draft an on-brand reply, and send it through the same tool.
Ask for a fan summary or a financial read across every connected creator.
Use it in Claude desktop, a Claude Project, or Claude Code with the same key.
Every tool call runs the same credit and rate-limit metering, tagged as MCP.
Calls go through signing and per-account proxies. Claude never sees raw credentials.
One key, one config block.
Generate a Fanduit key in the dashboard.
Drop the config block into Claude with your key.
"Which fans went quiet this week?" and Claude answers from live data.
Chat-ops for a whole roster.
Yesterday's revenue and top spenders across every account in one prompt.
Read a chat and draft a reply that matches the creator's voice.
Generate a financial and fan summary per creator on demand.
Ask which accounts need attention and why.
Run a whole roster from one API — connect every creator you operate and automate messaging, growth, and reporting at scale.
Neutral infrastructure on accounts you own, with encrypted credentials, per-account proxies, and no screen sharing.
Clean REST + MCP, signed webhooks, idempotency, and transparent credit metering. Build tools that don\'t break by morning.
Claude desktop, Claude Projects, and Claude Code all support MCP, so the same Fanduit key works across them.
No. Tool calls use the same credits and rate limits as the REST API, tagged as the MCP surface.
Yes, through the send-message tool. You decide how much autonomy to give it.
Add the MCP server and run your roster from Claude.