CallGrid

CALLGRID MCP / AUTOMATIONS

Your reports run themselves.

Connect CallGrid to Claude with the tools you already use, set the cadence, and a finished summary of any campaign, source, or destination lands in Slack, your inbox, a sheet, or Notion. No builder, no exports, no one logging in to run it.

Read only, so it is safe to share with your team. Works with the connectors you already run.

Live, hands-offA scheduled task running end to end in Claude Cowork
ClaudeCoworkSonnet

Scheduled

Daily revenue digest

Weekdays 7:00Next run

Weekly buyer scorecard

Mon 8:00in 3 days

Hourly anomaly watch

Hourlyin 42 min

Tasks run while the desktop app is open. A missed run retries on wake.

Daily revenue digestCallGridConnectedmcp.callgrid.com/mcp

Schedule

Every weekday

7:00 AM

CallGrid read

Used CallGrid · summary-report

240ms

Deliver

to Slack

#pay-per-call

Delivered to Slack

#pay-per-callRuns weekdays, 7:00 AM
CallGridAPPReports7:00 AM

Yesterday: 1,284 billable calls, $38,410 revenue, $29.91 RPC

Billable calls

1,284

6.2%

Revenue

$38,410

9.4%

RPC

$29.91

1.1%

TOP SOURCES BY RPC

  • Meridian Media$38.90
  • Vantage Media$32.60
  • Summit Leads$23.40
14 days agoYesterday

Posted to Slack by your Slack connector. CallGrid read only.

Sample data, anonymized demo. CallGrid reads only. Delivery is performed by your Slack, Gmail, Sheets, or Notion connector with your own permissions.

WHAT CHANGED

You used to ask. Now it just shows up.

The first version of CallGrid's MCP let you ask your data anything inside Claude or ChatGPT and get an answer with a chart. Now you can chain it. Connect CallGrid alongside Slack, Gmail, Sheets, or Notion, set a schedule, and the report you used to build by hand writes and delivers itself.

CallGrid is the source. Read only. It surfaces and analyzes your live numbers.

The assistant is the analyst. It reasons over the numbers and writes the summary.

The second connector is the delivery. It drops the finished work where the team already looks.

Connect in about 30 seconds

V1: you ask

“How did Medicare Inbound do yesterday?”

V2: it runs itself

Every weekday at 7:00 AM, posted to #pay-per-call

ScheduleCallGrid readSlack deliver

HOW THE CHAIN WORKS

Three pieces, one chain, on a cadence you set.

Cadence·Daily / Weekly / Monthly

Source

CallGrid

Reads your live data. Read only, never writes.

summary-report

Analyst

The assistant

Reasons over the numbers and writes the summary.

Claude · ChatGPT · any MCP client

Delivery

Your connector

Drops it where the team looks, with your permissions.

Slack · Gmail · Sheets · Notion

CallGrid never writes anywhere. The delivery connector does, with your own permissions in that tool.

FOUR WAYS TO DELIVER

Finished work, landing in the tools you already use.

Each of these is a real piece of work, building itself. CallGrid reads, the assistant writes, and the destination connector delivers. Revenue figures are estimates built from your own data, not guarantees.

Centerpiece ASample data, anonymized demo

Slack daily digest

Yesterday's billable calls, revenue, and RPC with day over day deltas, plus a top and bottom source leaderboard, posted to a channel every morning.

#pay-per-callRuns weekdays, 7:00 AM
CallGridAPPReports7:00 AM

Yesterday: 1,284 billable calls, $38,410 revenue, $29.91 RPC

Billable calls

1,284

6.2%

Revenue

$38,410

9.4%

RPC

$29.91

1.1%

TOP SOURCES BY RPC

  • Meridian Media$38.90
  • Vantage Media$32.60
  • Summit Leads$23.40
14 days agoYesterday

Posted to Slack by your Slack connector. CallGrid read only.

CallGrid read · summary-reportDeliver · Slack
Centerpiece BSample data, anonymized demo

Email scheduled report

A polished monthly buyer or partner performance summary that sends itself, the kind of report you used to export and format by hand.

Brightpath Insurance, May performance summary

CallGrid Reports <reports@brightpath-demo.com>

Sends on the 1st, 8:00 AM

Hi team, here is Brightpath Insurance's May performance across all buyers.

BuyerConnectsConn. rateConv.RevenueRPC
Cardinal Leads9,20471%3,118$284K$30.84
Summit Leads7,41568%2,406$211K$28.45
Apex Reach5,06659%1,402$132K$26.10

Revenue by campaign

$0K$100K$200K$300KMedicareACA Prem.Final Exp.U65

Cardinal Leads drove the strongest month on both volume and RPC. Apex Reach's connect rate slipped and is worth a routing review.

Sent by your Gmail connector with your permission. Revenue figures are estimates from your own data, not guarantees.

CallGrid read · summary-reportDeliver · Gmail
Centerpiece CSample data, anonymized demo

Sheets or Notion live sync

SheetsNotion

The running tab or page your team already opens stays current on its own. No export, no copy paste.

Performance, daily by sourceAppended daily by your Sheets connector
DateSourceCallsRPCRevenue
3May 12Vantage Media402$32.60$13,105
4May 13Summit Leads355$23.40$8,307
5May 14Meridian Media468$38.90$18,205

Or a Notion database, same data, updated on its own.

CallGrid read · summary-reportDeliver · Sheets
Centerpiece DSample data, anonymized demo

Lights-out cloud workflow

Zapiern8n

When the laptop is closed, the same CallGrid connection runs server side. Cowork is the cockpit, Zapier and n8n are the always-on backstop.

Zapier · cloud workflow

SCHEDULE7:00 AMCallGrid · MCPsummary-reportTHRESHOLDconn. rateALERTSlackDIGESTEmail

A schedule trigger fires the CallGrid read, a threshold branches, and the alert path posts to Slack. One token flows at a time.

SET IT UP IN WORDS, NOT CONFIG

Describe it once. It runs on its own.

Say what you want in plain English. The schedule, the CallGrid read, and the delivery are wired for you. Then manage it like any other task: pause it, run it now, or change the cadence.

Describe it once

“Every weekday at 7am, summarize yesterday's calls by source and post the top and bottom three to #pay-per-call”

Task created

Daily revenue digest

Weekdays 7:00CallGrid readSlack deliver

Scheduled tasks

4 active
Daily revenue digestDaily
Last run 7:00 AMSlack
Weekly buyer scorecardWeekly
Next run Mon 8:00Email
Hourly anomaly watchHourly
Last run 41m agoSlack
Cadence options:HourlyDailyWeekdaysWeekly

STRAIGHT TALK

Where the schedule runs, told straight.

Two honest options. Pick the cockpit when you want to watch and tweak the report. Pick the backstop when it has to land whether or not a laptop is open.

The cockpit

In Claude Cowork

  • Runs on your Mac while the desktop app is open and the computer is awake.
  • You watch it build the report and can tweak it on the spot.
  • A missed run retries when the machine wakes.

Best for the reports you want to see and tweak.

The backstop

In the cloud

  • The same CallGrid connection runs server side through Zapier or n8n as an MCP client.
  • A 7:00 AM report lands whether or not a laptop is open.
  • Threshold alerts can fire the moment something moves.

Best for can't-miss deliveries and instant alerts.

THRESHOLD AND ANOMALY ALERTS

Not just the daily digest.

Event-driven alerts can fire the moment a connect rate drops or a source's bid rejections spike, by watching CallGrid's webhooks, pushed straight to Slack or email.

#alerts
CallGridALERTnow

Connect rate on Medicare Inbound dropped 18% vs the 7-day average

7-day avg 64%now 52%
View in CallGrid

Event-driven alerts fire the moment something moves, not just on the daily schedule. Sample data, anonymized demo.

PROMPTS AND RECIPES TO STEAL

Copy a recipe. Set the cadence. Done.

Real scheduled use cases. Each shows its chain as three small chips: the schedule, the CallGrid read, and where it delivers.

  • Daily revenue digest

    Every weekday at 7am, summarize yesterday’s billable calls, revenue, and RPC with day-over-day deltas, plus the top and bottom three sources, and post it to #pay-per-call.

    Weekdays 7:00CallGrid readSlack
  • Monthly buyer scorecard

    On the 1st of each month, build a buyer performance summary for last month with connects, connect rate, conversions, revenue, and RPC per buyer, and email it to the partners list.

    Monthly, 1stCallGrid readEmail
  • Source leaderboard sync

    Every day at midnight, append yesterday’s per-source calls, RPC, and revenue as a new row to the Performance sheet.

    DailyCallGrid readSheets
  • Campaign status board

    Every Monday, refresh the Campaign Performance database in Notion with each live campaign’s RPC and revenue for the prior week.

    Weekly, MonCallGrid readNotion
  • Connect-rate watch

    Every hour, check connect rate by source against the 7-day average and post an alert to #alerts if any source drops more than 15%.

    HourlyCallGrid readSlack
  • Rejection spike alert

    Watch bid rejections by buyer. If a buyer’s rejection rate doubles versus its trailing average, email me the buyer, the reason breakdown, and the affected campaigns.

    Event-drivenCallGrid readEmail

READ ONLY BY DESIGN

Safe to schedule. Safe to share.

The CallGrid side surfaces and analyzes your data. It does not place calls, create or edit campaigns, move numbers, or change settings. The delivery into Slack, Gmail, Sheets, or Notion is done by those connectors, each with its own permissions you control. So you can set these to run and hand the outputs to anyone who needs answers but should not be editing the account.

Reads your live data

The same numbers you see in CallGrid Reporting.

Never writes to your account

No placing calls, no editing campaigns, no moving numbers.

Delivery uses your permissions

Each connector posts with the access you control in that tool.

WORKS WITH

Claude, ChatGPT, and the tools you already use.

  • Claude
  • ChatGPT
  • Slack
  • Gmail
  • Sheets
  • Notion
  • Zapier
  • n8n

Server URL

mcp.callgrid.com/mcp

No keys, authorize once

Connect in 30 seconds

  1. 1

    Add the URL

    Paste mcp.callgrid.com/mcp as a connector in Claude.

  2. 2

    Authorize

    Sign in to CallGrid once. No keys to manage.

  3. Connected

    CallGrid is in your connector slot, ready to schedule.

WHERE IT FITS

One live dataset, behind every channel.

The MCP reads the same live data as CallGrid Reporting, Real-Time Bidding, and Call Routing, and the same webhooks and postbacks power the event-driven alerts.

FAQ

Questions, answered straight.

  • A recurring summary of your CallGrid data that an AI assistant writes and a delivery connector drops into Slack, Gmail, Google Sheets, or Notion on a cadence you set. CallGrid is the read-only source, the assistant is the analyst, and the second connector does the delivery.

  • Slack, Gmail, Google Sheets, and Notion today, through their own connectors, plus the cloud automation platforms Zapier and n8n. It works with Claude, ChatGPT, and any MCP client.

  • In Claude Cowork a scheduled task runs on your own machine while the desktop app is open and the computer is awake, and a missed run retries when the machine wakes. For delivery that has to land whether or not a laptop is open, you run the same CallGrid connection server side in the cloud through Zapier or n8n.

  • No. CallGrid’s MCP is read only. It surfaces and analyzes your data. It never places a call, creates or edits a campaign, moves a number, or changes a setting.

  • The delivery into Slack, Gmail, Google Sheets, or Notion is done by those connectors, each acting with the permissions you control in that tool. Because the CallGrid side is read only, you can hand the outputs to anyone who needs answers but should not be editing the account.

  • No. You describe the report once in plain English and the schedule, the CallGrid read, and the delivery are wired for you. The cloud path through Zapier or n8n is optional, for can’t-miss deliveries when the laptop is closed.

  • Hourly, daily, on weekdays, weekly, or monthly, per destination, source, or campaign.

  • The hosted MCP server is available to CallGrid customers. Connect your account, authorize once, and schedule.

Stop running the report. Let it run itself.