Jomablue MCP: Tokens and Usage
Jomablue MCP: Tokens and Usage
What Are Tokens?
Knowing what uses your AI plan’s allowance helps you get more answers from it. Every question you ask through Jomablue MCP uses tokens, and a few simple habits keep usage down.
AI clients measure usage in tokens. A token is a small chunk of text, about 3.5 characters. Your AI plan’s usage limit is counted in tokens, even if your AI client never shows you a number.
- Everything counts, in both directions. Your AI client uses tokens for what it reads (your question, the chat so far, and the data Jomablue sends back) and for what it writes (the answer).
- The data Jomablue sends back is usually the biggest part. A short question can bring back a lot of event data.
- Long chats cost more. Your AI client re-reads the whole chat every time you send a message, so the tenth question in a chat costs more than the first.
What Uses the Most
| What your AI client reads | Approximate tokens |
|---|---|
| The full HTML of one email or page | 5,000 to 75,000 or more |
| A list of 25 campaigns | 7,500 |
| Each connected Organisation (fixed) | 6,000 to 8,000 |
| A list of 25 vendors | 4,500 |
| A list of 25 templates | 2,500 |
| One event’s details, or one email template’s text | 1,000 to 3,000 |
Ways to Keep Usage Down
- Ask about one event at a time unless you need to compare events.
- Ask for the wording, not the full HTML. The text of an email template is far smaller than its HTML.
- Ask for a specific check or a summary, not for everything at once.
- Start a new chat for a new event or topic. This stops your AI client re-reading an old conversation.
- Connect only the Organisations you need. Each one adds its fixed amount.
How to Estimate Usage
Work out one request
- Lists come back 25 items at a time by default, and up to 100.
- Each request costs a small fixed amount plus an amount for each item: tokens for one request = fixed amount + (number of items x amount per item).
- For example, a list of 25 templates is about 50 + (25 x 100), or about 2,550 tokens.
Plan a rollout
- Try the real task on one event.
- See how much it used.
- Multiply up, then add a buffer. Expect a range, not a single number.
Usage Figures
These are Jomablue’s estimates for the data sent back, based on about 3.5 characters per token. Actual figures vary with your event data. They don’t include your question, the answer your AI client writes, or anything else in the chat.
Lists
| What the AI client reads | Fixed | Per item | 25 items | 100 items |
|---|---|---|---|---|
| Events | 50 | 110 per event | 2,800 | 11,000 |
| Pages | 50 | 150 per page | 3,800 | 15,000 |
| Categories | 50 | 80 per category | 2,000 | 8,000 |
| Registration admissions, stores or embeds | 30 | 60 per item | 1,500 | 6,000 |
| Registration field search | 80 | 60 per field found | 1,600 | 6,000 |
| Campaigns | 60 | 300 per campaign, more with complex audiences | 7,500 | 30,000 |
| Templates | 50 | 100 per template | 2,500 | 10,000 |
| Merge tag check on templates | 150 | 140 per template | 3,600 | 14,000 |
| Vendors | 50 | 180 per vendor, more with many traits | 4,500 | 18,000 |
| Vendor leads report | 40 | 80 per vendor, plus 40 per sponsored session | 2,000 or more | 8,000 or more |
| Session attendance report | 60 | 130 per session | 3,300 | 13,000 |
| Registration counts by category | 30 | 100 per category | 2,500 | 10,000 |
| Program member counts | 60 | 150 per program | 3,800 | 15,000 |
| Program status breakdown | 60 | 30 per status | 800 | 3,000 |
Single items and fixed cost
| What the AI client reads | Approximate tokens |
|---|---|
| Each connected Organisation (fixed cost) | 6,000 to 8,000 |
| Access status check | 1,500 or more |
| Event details | 1,000 to 3,000 |
| Registration store | 1,000 to 4,000 |
| Reference document, such as the Campaigns or Pages reference | 3,000 each |
| Email template (text) | 1,000 to 3,000 |
| Email or page HTML body | 5,000 to 75,000 or more |
Additional Notes
- Usage limits. Usage limits are set by your AI client’s provider. They differ by provider and plan, so check your AI client’s help pages for yours.
- Estimated, not measured. The program figures and the fixed cost per Organisation are estimated, not measured. The other figures were measured, and all of them vary with your event data.
Frequently asked questions
Does Jomablue set my usage limit?
No. Usage limits are set by your AI client’s provider and depend on your plan.
Why did one question use so much?
The two most common reasons are a request that reads full email or page HTML, and a long chat that your AI client has to re-read with every message.