Skip to main content

Usage Limits

Understanding how PlexMCP counts and limits usage across your organization.

What Counts as a Request?

A request is counted each time you call the PlexMCP API to interact with an MCP. This includes:

  • Tool invocations: Calling a tool on an MCP
  • Resource reads: Reading resources from an MCP
  • Prompt completions: Getting prompts from an MCP
  • Connection tests: Testing MCP connectivity from dashboard

What Doesn't Count

  • Dashboard page views
  • API key management
  • Team management
  • Viewing analytics
  • Health check polling (automatic)

Plan Limits

Request Limits

PlanMonthly Requests
Free1,000
Pro50,000
Team500,000
EnterpriseCustom

MCP Limits

PlanMax MCPs
Free5
Pro15
TeamUnlimited

Team Member Limits

PlanMax Members
Free1
Pro5
TeamUnlimited

API Connection Limits

PlanMax API Connections
Free2
Pro10
TeamUnlimited

Monitoring Usage

Dashboard View

  1. Go to Dashboard or Billing
  2. See current usage vs. limits
  3. View usage breakdown by day

Usage by MCP

See which MCPs consume the most:

  1. Go to MCPs
  2. View request counts per MCP
  3. Sort by usage

Usage by API Key

Track usage per key:

  1. Go to API Keys
  2. View request counts per key
  3. Identify high-usage keys

Usage Alerts

Configure notifications when approaching limits:

  1. Go to BillingAlerts
  2. Set threshold (e.g., 80%)
  3. Choose notification method
  4. Save settings

Alert types:

  • Warning (80%): Time to review usage
  • Critical (95%): Consider upgrading
  • Exceeded (100%): Action required

What Happens at Limits

Request Limits

Free Plan:

  • API calls return 429 Too Many Requests
  • Service resumes at next billing period
  • Can upgrade to continue immediately

Paid Plans:

  • Overage charges apply ($0.001/request)
  • No service interruption
  • Invoice adjusted at period end

MCP Limits

When at MCP limit:

  • Cannot add new MCPs
  • Existing MCPs continue working
  • Must delete or upgrade to add more

Team Member Limits

When at member limit:

  • Cannot send new invites
  • Existing members unaffected
  • Remove member or upgrade to invite more

API Key Limits

When at key limit:

  • Cannot create new keys
  • Existing keys work normally
  • Revoke unused keys or upgrade

Optimizing Usage

Reduce Unnecessary Calls

  1. Cache responses: Store results when appropriate
  2. Batch operations: Combine related calls
  3. Use webhooks: React to events vs. polling

Monitor for Issues

  1. Check error rates: Failed calls still count
  2. Review retry logic: Avoid excessive retries
  3. Audit API keys: Find and fix runaway integrations

Right-size Your Plan

If consistently near limits:

  • Upgrade to next tier
  • Consider annual billing for savings
  • Contact sales for custom limits

If consistently under limits:

  • Review if you need current tier
  • Downgrade to save costs

Rate Limiting

In addition to monthly limits, PlexMCP enforces rate limits:

PlanRequests/second
Free10
Pro100
Team1,000
EnterpriseCustom

Rate limit responses:

  • Status: 429 Too Many Requests
  • Header: Retry-After: <seconds>
  • Body: Error message with wait time

Handling Rate Limits

async function callWithRetry(fn, maxRetries = 3) {
for (let i = 0; i < maxRetries; i++) {
try {
return await fn();
} catch (error) {
if (error.status === 429) {
const waitTime = error.headers['retry-after'] || 1;
await sleep(waitTime * 1000);
} else {
throw error;
}
}
}
throw new Error('Max retries exceeded');
}

Billing Period

  • Start: Day of signup/upgrade
  • End: Same day next month
  • Reset: Limits reset at period start

Example:

  • Signed up: January 15
  • Billing date: 15th of each month
  • Usage resets: Every 15th

FAQs

Do failed requests count?

Yes, all API calls count toward limits, including errors.

Can I prepay for requests?

Contact sales for custom request packages.

What's the grace period?

None on Free plan. Paid plans have automatic overage.

Can I set hard limits?

Yes, contact support to disable overages (paid plans).

How quickly does usage update?

Near real-time in dashboard, within ~1 minute.