promtpress-mcp
MCP server for PromtPress — buy Telegram subscribers and website traffic from your AI assistant. The agent spends within a mandate; it can never fund or withdraw.
Documentation
PromtPress Ads — MCP server
Buy advertising from your AI assistant. PromtPress delivers outcomes —
Telegram channel subscribers and website traffic — billed per verified result
from a single prepaid balance.
Endpoint: `https://mcp.promtpress.com/mcp` (Streamable HTTP)
Auth: OAuth 2.1 — dynamic client registration, PKCE required
Docs: https://promtpress.com/mcp
Why this exists
Every ad platform now has an MCP server, and they all do the same thing: let an
agent drive a dashboard you already had to set up. You still need an account per
network, a payment method per network, ad review per network, and creatives per
network.
This one is different in the only way that matters for an agent: **it sells the
outcome, not the dashboard.** Pick a placement format and a goal, get a forecast,
launch. Media buying, source mix, fraud filtering and billing happen on our side.
The safety property
The agent can spend, but it cannot fund and it cannot withdraw.
There is no tool that adds money, changes payout details, or moves funds
anywhere. Top-ups happen only in the browser, only by the account owner. A fully
compromised agent — leaked token, prompt injection, whatever — can at most spend
what is already on the balance, within the mandate the owner set.
That is the design constraint the whole server is built around. Everything else
is a consequence of it.
Guardrails
| Control | Default |
|---|---|
| Per-operation cap | $50 |
| Daily cap | $100 |
| Monthly cap | set by owner |
| First campaign on a new account | always confirmed by a human |
| Anything above the mandate | confirmation link to the owner, expires in 15 min |
| Mandate | can be paused instantly by the owner |
Confirmation links carry a fingerprint of the parameters shown to the human.
Change the budget after the person approved it, and the approval no longer
matches — it burns instead of executing.
Hard caps are enforced server-side: no mandate can be raised above $1,000 per
operation or $5,000 per day, whatever the owner sets.
Tools
Read — `get_account`, `list_formats`, `analyze_channel`, `list_campaigns`,
`get_campaign`, `get_stats`, `get_ledger`, `get_topup_link`
Write — `signup`, `estimate`, `create_campaign`, `update_campaign`,
`submit_campaign`, `set_budget`, `pause_campaign`, `resume_campaign`
`submit_campaign` is the only tool that spends money. Everything else either
reads state or prepares a draft.
Connect
Claude — Settings → Connectors → Add custom connector →
`https://mcp.promtpress.com/mcp`
Claude Code
claude mcp add --transport http promtpress https://mcp.promtpress.com/mcpAnything else that speaks MCP — point it at the endpoint above. The server
advertises its OAuth metadata at
`/.well-known/oauth-protected-resource`, so clients that support OAuth 2.1 will
discover and register themselves.
On first use you will be asked for the email of your PromtPress account and sent
a one-time confirmation link. No account yet? The `signup` tool creates one — the
agent can do that part.
What we do not sell
Gambling, betting, adult content, and anything requiring a licence we cannot
verify. Campaigns in these categories are rejected before they reach a source.
Support
Frequently asked questions
What is promtpress-mcp?
promtpress-mcp is MCP server for PromtPress — buy Telegram subscribers and website traffic from your AI assistant. The agent spends within a mandate; it can never fund or withdraw.
How do I install promtpress-mcp?
Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.
Is promtpress-mcp open source?
Yes — it is hosted on GitHub at https://github.com/x3clv/promtpress-mcp.
Related MCP tools
Model Context Protocol Servers
Pre-indexed code knowledge graph, auto syncs on code changes, for Claude Code, Codex, Gemini, Cursor, OpenCode, AntiGravity, Kiro, CoPilot, and Hermes Agent — fewer tokens, fewer tool calls, 100% local
an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM
The easy-to-use open source Business Intelligence and Embedded Analytics tool that lets everyone work with data :bar_chart:
The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra
Cognee is the open-source AI memory platform for agents. Give your AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.
Run your own MCP server? See who uses it and what to fix.
Measure it with TrackMCP