For AI assistants

The agent-ready operating system for music PR agencies

Your AI assistant reads your campaign memory, drafts your next actions, and queues work for your approval. It does not scrape spreadsheets or touch your relationships directly.

TAP drafts and queues. It never sends on its own. You approve every send.

Connect in 5 minutes Full technical reference

50 tools across 12 capability areas. 28 REST endpoints. All human-gated.

What your agent can do

Plain-language capabilities

These are the things you can ask your assistant to do. It reads your workspace, makes suggestions, and writes drafts. It does not take action until you confirm.

Find and enrich contacts

Your assistant can search your contact database, pull up relationship history, and run enrichment to fill in missing details like outlet, genre, and response rate.

Draft pitches

Hand it a campaign and a list of contacts. It writes tailored pitch drafts for each one and puts them in your queue for review. Nothing leaves until you say so.

Run campaigns

Create campaigns, add contacts, log plays and replies, and keep everything in one place. Your assistant can read all of this context before it does anything.

Queue work for your approval

Every draft, every proposed send, every suggested action lands in your approval queue first. The assistant proposes; you decide.

Build client reports

Once outcomes are logged, your assistant can pull a summary across plays, replies, and coverage and assemble the bones of a client report.

Read campaign memory

TAP stores what has happened on every campaign. Your assistant can read that history before it drafts anything, so its suggestions are grounded in the actual record.

Capabilities are grounded in 50 MCP tools and 28 REST endpoints. The full technical list lives on /developers.

How it works in practice

Three scenes from a real campaign week

Scene 1

Your morning brief

Overnight, your assistant catches up so you do not have to

While you sleep, radio plays come in, replies land, and the campaign moves. In the morning your assistant reads all of it, works out who still needs a follow-up, drafts the messages, and parks them in your approval queue. You open TAP, see what is ready, and decide what goes out.

Scene 2

Drop a list, get pitches

Hand over a contact list and come back to polished drafts

Paste in a list of contacts or import a spreadsheet. Your assistant checks them against your existing database, fills in missing details where it can, and drafts a tailored pitch for each one based on the campaign brief. You review the drafts, adjust anything you want to change, and approve the ones ready to send.

Scene 3

Log outcomes, get the report

As plays and replies come in, the report builds itself

Log a WARM play, note a positive reply, record a coverage mention. Your assistant tracks all of it against the campaign and, when you need a client update, assembles the numbers and narrative from the actual record rather than asking you to remember what happened.

The one thing that never changes

TAP drafts and queues. It never sends on its own.

TAP can analyse, suggest, draft, and queue. TAP never auto-sends without explicit human approval.

Every pitch draft, every proposed follow-up, every suggested send lands in your approval queue first. You review it, adjust it if needed, and press send yourself. Your relationships stay yours.

Three ways to connect

Pick the one that fits how you work

Through your AI assistant

If you use Claude Desktop, Cursor, or a custom agent setup, install the TAP MCP server once and your assistant can talk to TAP directly. Best if you already have an AI assistant you trust and want to extend it.

From the command line

The TAP CLI gives you quick access to contacts, campaigns, and skills without opening a browser. Useful for batch operations or if you prefer working in a terminal.

Via the API

Every action in TAP is available through the REST API. Right choice if you are building a custom integration or want to connect TAP to another tool in your stack.

Connect in 5 minutes

From zero to first command

This is the setup for connecting Claude Desktop, Cursor, or any MCP-compatible assistant. The API and CLI routes are on /developers.

1
Copy your API key
Go to Settings then API Keys inside TAP and create a new key. Give it a name like "My assistant" and choose the scopes you need. Copy the key now -- it is only shown once.
2
Install the MCP server
The TAP MCP server runs on-demand via npx -- nothing to install globally unless you prefer it.
Terminal
```
npx -y @totalaudiopromo/tap-mcp@latest
```
3
Paste the config into your assistant
Add the block below to your assistant config file. On macOS for Claude Desktop that is ~/Library/Application Support/Claude/claude_desktop_config.json. Replace the placeholder key with your real one.
claude_desktop_config.json
```
{
  "mcpServers": {
    "tap": {
      "command": "npx",
      "args": ["-y", "@totalaudiopromo/tap-mcp@latest"],
      "env": {
        "TAP_API_KEY": "tap_ak_live_..."
      }
    }
  }
}
```
4
Run your first command
Restart your assistant. Then ask it something like:
“List my active campaigns and draft tomorrow's follow-ups.”
TAP will return your campaigns, check reply and play status, and put draft follow-ups in your approval queue.

Or use the REST API directly

curl https://totalaudiopromo.com/api/v1/campaigns \
  -H "Authorization: Bearer $TAP_API_KEY"

Full API reference and all available scopes are on /developers.