Connect your AI to ViewsMax

ViewsMax exposes an MCP server and a REST API so AI assistants can post to your connected social accounts, manage offers and tracked links, and read your analytics — on your behalf, with your permission. Machine-readable version of this page: /ai.md. New to MCP? Start with the MCP server overview — what it exposes, how auth works, and every tool grouped by job.

Pick your AI agent

Endpoints

Credentials

Two options, both sent as a Bearer token:

  1. OAuth (recommended for chat apps). Point an MCP client at the endpoint; sign in and approve in the browser. The consent screen offers read-only or full access. No key handling.
  2. API key (for headless agents and scripts). Settings → AI Assistant Access → generate key (vmx_...). Shown once; rotating invalidates the old one. The same key works on the REST API (posts, offers, tracking, stats — read-only keys are limited to GET).

Claude (claude.ai / Desktop)

Customize → Connectors → + → Add custom connector → paste the MCP endpoint → Add → Connect and complete the sign-in approval. The connector also works in the mobile apps and Cowork.

https://api.viewsmax.com/api/mcp

Full Claude guide →

Claude Code

claude mcp add --transport http --scope user viewsmax https://api.viewsmax.com/api/mcp

or in .mcp.json:

{ "mcpServers": { "viewsmax": { "type": "http", "url": "https://api.viewsmax.com/api/mcp" } } }

Then run /mcp inside Claude Code and sign in.

Full Claude Code guide →

ChatGPT

Settings → Security and login → turn on Developer mode. Then go to chatgpt.com/plugins → + → paste the MCP URL and complete OAuth. ViewsMax is an action connector (create/schedule posts, read stats) — use it from regular chats.

Full ChatGPT guide →

Codex

codex mcp add viewsmax --url https://api.viewsmax.com/api/mcp
codex mcp login viewsmax

Full Codex guide →

Cursor

.cursor/mcp.json:

{ "mcpServers": { "viewsmax": { "url": "https://api.viewsmax.com/api/mcp", "headers": { "Authorization": "Bearer vmx_YOUR_KEY" } } } }

Full Cursor guide →

OpenClaw

Download the ViewsMax skill from /skills/viewsmax/SKILL.md into your agent's skills directory and set VIEWSMAX_API_KEY in its environment. The skill drives the REST API; alternatively point OpenClaw's MCP support at the endpoint above.

Full OpenClaw guide →

Hermes Agent

~/.hermes/config.yaml:

mcp_servers:
  viewsmax:
    url: "https://api.viewsmax.com/api/mcp"
    auth: oauth

Full Hermes Agent guide →

Plain REST / curl

curl -H "Authorization: Bearer vmx_YOUR_KEY" https://api.viewsmax.com/api/posts

Responses use a { success, message, data } envelope. Full reference: https://api.viewsmax.com/docs.

What agents can do (30 MCP tools)

list_connected_accountslist_brandsupload_mediacreate_postlist_postsget_postupdate_postdelete_postlist_offerscreate_offerget_offerupdate_offerdelete_offercreate_tracking_linkget_offer_statsget_stats_timeseriesdisconnect_accountget_connect_urlcreate_feature_requestlist_outlierssearch_outliersget_outlierfetch_outlierget_outlier_breakdowngenerate_outlier_breakdownlist_saved_outlierssave_outlierremove_saved_outlieradd_outlier_channelget_outlier_channel_ingest

Typical posting flow: list_connected_accounts → upload_media → create_post (draft / posted / scheduled) → publishing is asynchronous, so poll get_post for per-platform results. The same tools grouped by job, with what each needs: /mcp.

Security & limits

  • Read-only credentials cannot write, anywhere.
  • Every AI tool call is recorded in your audit log (Settings → AI Assistant Access).
  • Rate limits: 120 MCP requests/min per token; 180 create_post/hour; 40 upload_media/hour.
  • Rotate your API key any time to revoke access instantly.