Pay as you go
Start with 50 free MCP calls every month. Paid plans auto-renew every 30 days, with optional usage-based overage billing.
How subscriptions work
- Every account gets 50 free MCP calls per calendar month, resetting on the 1st.
- Paid MCP request plans are subscriptions and auto-renew every 30 days unless canceled.
- If usage-based fees are enabled, overage usage is billed at $0.10 per request up to your budget cap.
- Usage-based fees are charged on renewal as one consolidated line item for the previous cycle.
Free Forever
Everything you need to get started
Need more? Choose a subscription.
MCP request and analytics plans are subscriptions that auto-renew every 30 days. After the included sources, each additional source is $3.
Extra Source
Add another documentation URL, repo, or website to a specific MCP server
per additional source
First 2 sources are included. After that, each additional source is $3.
100 Requests / month
Starter subscription
100 MCP requests per month
Up to 1M indexed chunks per server
Auto-renews every 30 days
300 Requests / month
Growth subscription
300 MCP requests per month
Up to 2M indexed chunks per server
Auto-renews every 30 days
1,000 Requests / month
Scale subscription
1,000 MCP requests per month
Up to 3M indexed chunks per server
Auto-renews every 30 days · Highest MCP request tier
Dashboard Analytics Subscriptions
Each tier includes everything in the tier before it. Telemetry for the Action and Predictive tabs is captured on every request no matter which tier you are on, so upgrading fills those tabs with your existing history. Full walkthrough in the analytics docs.
Core Analytics
$99/mo
How much your MCP server is used.
- Totals, success rate, latency, active sources
- Requests over time, tool, source, and client charts
- Recent request log with the end-user query
- CSV export on every chart
Core + Action Tab
$149/mo
See exactly which passage answered which question.
- The exact text sent to the AI, with page and section
- Code snippet provenance back to source, page, section
- Per-page retrieval confidence and content gaps
- Impact model with the formula shown in full
Core + Action + Predictive
$199/mo
Know what an answer costs — and how to cut it.
- MCP requests to a working answer, per question chain
- A–D quality grade for every source
- Ranked content fixes with projected request counts
- 30-day volume, requests avoided, and value forecasts
Common questions
What counts as an MCP call?
Each tool invocation from an MCP client (like Cursor or Claude Desktop) counts as one call. Search, code extraction, Q&A — each is one call.
How do monthly MCP requests work?
Plans renew every 30 days with their included request allowance. Overage usage can continue at $0.10/request if you enable usage-based fees and set a monthly budget cap.
What if I upgrade plans mid-cycle?
Requests are now subscription-based, so upgrades change your active plan for the next billing cycle and renew automatically every 30 days unless canceled.
Are extra sources permanent?
No. The first 2 sources per server are included for free. After that, each additional source is $3.
How many MCP servers can I create?
Every account can create up to 3 MCP servers for free. Each server includes 2 sources. Need more than 3 servers? Contact zair@appatools.com for an Enterprise plan.
What happens when I hit my call limit?
MCP calls will return a friendly limit message unless usage-based fees are enabled. Your servers stay online and your data is never deleted.
What is the highest request plan?
The highest MCP plan includes 1,000 requests per month. For higher usage, enable usage-based fees to keep serving requests over your included plan amount.
Is this a subscription?
Yes. Paid plans are subscriptions that automatically renew every 30 days unless canceled.
How large can my sources be?
Source size is measured in indexed chunks — the small passages we split your content into so an AI agent can retrieve exactly the right part. The allowance is per MCP server and cumulative: all of that server's sources share it. Free includes up to 500K chunks per server, and each paid request plan raises it, up to 3M on the highest plan.
What happens if a source is too large for my plan?
Indexing stops at your allowance instead of failing. Everything already indexed stays queryable and your MCP server keeps working. You can upgrade for a larger allowance, or remove another source from that server to make room.