Rovo MCP Pricing: Free Calls, Metered Tools, and a 10-Person Team Budget

Rovo MCP is free to install and connect, but enriched-context workflows consume credits. For a hypothetical 10-person team with Jira Standard and Confluence Standard, our baseline projects $15, $95, or $195 in monthly overages at one, five, or ten enriched prompts per person daily once extra-usage billing begins. Those figures assume 20 working days, 10 credits per prompt, and no competing Rovo consumption.

As of September 2026, Atlassian lists Rovo MCP installation and connection at $0. Single-product lookups and updates are also free, while enriched context and unified search consume credits. Identify the operation behind each request before treating an external assistant’s workflow as free. Sources: Atlassian’s credit rules and licensing FAQ.

By Remote Work Picks
Pricing and features as of September 2026 · Reviewed September 6, 2026

We compared documented features and pricing, then calculated the scenarios below. This is a budgeting model, not measured customer usage or a guarantee of future charges. The links in this article are reference and internal editorial links; no affiliate purchase links are included.

Is Atlassian Rovo MCP free? What changes on December 3, 2026

Separate three questions: whether you can connect, whether an operation consumes credits, and whether that consumption creates an extra bill.

MCP provides a connection between your external AI client and Atlassian Cloud. The client can use that connection to retrieve information or perform actions within your existing permissions. The connector itself does not make every workflow free. Source: Atlassian’s developer overview.

Atlassian says allowance limits and extra-usage billing for Rovo credits take effect on December 3, 2026. A forecast made in September therefore estimates spending under the announced billing model; it is not a September overage invoice. Source: Atlassian’s pricing announcement.

For a distributed team, the useful distinction is between retrieving a known item and asking an assistant to assemble context. “Open this planning page” and “explain what changed across our launch work” may sound equally routine during a morning catch-up. Budget them according to the operations the assistant actually invokes.

Free versus metered calls

The following comparison reflects documented treatment as of September 2026. Examples describe operations, rather than promising that particular wording will force an assistant to choose a free route.

Sources: Atlassian’s credit classification and MCP usage documentation.

Atlassian’s MCP usage documentation identifies search, searchAtlassian, and getTeamworkGraphContext as credit-consuming tools. Its developer overview says Teamwork Graph and Rovo Search tools consume up to 10 credits per call. That limit concerns one call; an assistant’s multi-call prompt can consume more. Sources: MCP usage documentation and developer overview.

Use that per-call statement narrowly. Atlassian’s broader credit rules describe variable pricing, with most enriched calls using 1–10 credits. They do not establish a universal ten-credit maximum for every possible enriched operation. Source: How Rovo credits work.

Our practical preference is to provide an issue key or page link when the task already has a known source. Broader context is worth evaluating when someone needs to discover connections, resolve conflicting updates, or prepare an async handoff. The tradeoff is more variable consumption, which a simple prompt count cannot fully explain.

How ten people build a shared credit pool

For this worksheet, assume the same ten people each have one Jira Standard seat and one Confluence Standard seat, with both subscriptions in the same Atlassian organization.

Atlassian lists 25 credits per user per month for each Standard subscription, with allowances combined across apps at the organization level. Source: Rovo licensing.

This is a hypothetical subscription mix. Check your actual allowance before copying 500 into a forecast.

The pool is shared with credit-consuming Rovo Chat, Studio, Agents, and Teamwork Graph activity. It is not an MCP-only allocation. Source: Atlassian’s MCP usage documentation.

That has an immediate planning consequence: dividing 500 by ten produces a useful average, but does not reserve 50 credits for each colleague.

At our assumed ten credits per enriched prompt, the entire pool covers 50 prompts per month. Spread evenly across ten people, that is five prompts each per month. One enriched prompt per person every working day already exceeds this baseline allowance.

Do not buy extra seats simply to improve that number. One additional Standard seat contributes 25 credits, equivalent to only $0.25 of modeled overage capacity at the published rate. Any subscription change needs to justify its full cost and useful features.

Monthly budget scenarios: light, regular, and heavy usage

These scenarios use the following inputs:

The rate is Atlassian’s published list price as of September 2026. Source: How Rovo credits work.

Calculate consumption first:

Monthly enriched prompts
= people × working days × daily enriched prompts per person

Monthly MCP credits
= monthly enriched prompts × average credits per prompt

Projected monthly overage
= max(0, MCP credits + other Rovo credits − included credits) × $0.01

Baseline: ten credits per prompt

For the regular scenario:

10 people × 20 days × 5 prompts = 1,000 prompts
1,000 prompts × 10 credits = 10,000 credits
(10,000 − 500) × $0.01 = $95

These are full-month planning scenarios after billing starts. They are not an estimate of the first December invoice, whose covered period needs checking.

“Light,” “regular,” and “heavy” are labels for our chosen frequencies, not observed customer categories. A project lead preparing broad handoffs could consume more credits with fewer prompts than someone retrieving individual work items.

Sensitivity: 25 credits per prompt

Keep everything else unchanged and increase the assumed average to 25 credits.

The regular forecast rises from $95 to $245, despite identical headcount and prompt frequency.

That $150 difference is why measuring credits per prompt matters more than refining a 20-day calendar assumption. Ten and 25 credits are planning inputs here. Neither is a guaranteed charge, and 25 is not a worst-case ceiling.

Add competing Rovo usage

Suppose the regular baseline also includes 3,000 other Rovo credits during the month:

MCP credits:                       10,000
Other Rovo credits:                 3,000
Combined consumption:             13,000
Included allowance:                  500
Projected overage: 12,500 × $0.01 = $125

The total becomes $125, compared with $95 when other consumption was zero.

For departmental accounting, avoid assigning the full 500-credit allowance to both MCP and another Rovo workflow. The organization receives that allowance once.

If you need the incremental cost attributable to MCP, subtract the forecast without MCP from the forecast with it. In this example, other usage alone would produce $25 in overages; adding MCP raises the total to $125, an incremental $100.

Copyable spreadsheet worksheet

Put these labels in column A and values or formulas in column B. Change B4 to 1 or 10 for the other frequencies; change B5 to 25 for the sensitivity case.

Format B13 as currency. The default inputs should return $95.

Why one prompt can cost several calls

Consider this hypothetical request: “Prepare tomorrow’s handoff: find the launch blockers, connect them to the planning decisions, and update the tracking item.”

An illustrative execution could look like this:

The numbers are invented solely to demonstrate the arithmetic. They are not tool tariffs or a recorded execution. The external assistant may choose a different sequence.

For measured budgeting, calculate:

Average MCP credits per prompt
= credits attributable to sampled MCP workflows
  ÷ number of sampled prompts

Include all the calls associated with those prompts. Record retries and follow-up searches so you can reconcile their actual treatment against usage records; do not assume every attempted call is billed.

For recurring workflows, a more useful forecast is:

Monthly MCP credits
= sum of (monthly runs of each workflow × its average credits per run)

This lets you distinguish a short daily lookup from a weekly project review without pretending they have the same cost.

What the estimate excludes

The worksheet calculates Rovo overages only. It excludes:

• Jira and Confluence base subscriptions.
• External AI-client subscriptions, API tokens, or other client charges.
• Taxes, currency conversion, and negotiated commercial terms.
• Prepaid-pack discounts.
• Other separately billed services or meters used by the workflow.

Include those expenses separately when building your remote-team software stack budget.

The main operational tradeoff also sits outside the dollar estimate: a low spending limit protects the budget but may leave someone unable to run a context-heavy handoff near month-end.

Before December 3: measure usage, set limits, and choose your cutoff

Rovo extra-usage billing begins December 3, 2026, at a published list rate of $0.01 per credit beyond the included allowance. Atlassian says extra usage is enabled by default; administrators can disable it or set spending limits. September forecasts prepare for future billing. Sources: pricing announcement and credit rules.

1. Confirm the allowance in your organization. Open Atlassian Administration → Insights → Platform usage → Rovo credits. Check the total and export usage data for analysis. Source: Atlassian’s usage reporting instructions.

2. Sample a representative workweek. Include ordinary lookups, async handoffs, and one broader planning task. Record the client, prompt count, tool calls where visible, and attributable credits. If the export cannot isolate a workflow, label the allocation as an estimate.

3. Replace both consumption assumptions. Update average MCP credits per prompt and other Rovo credits. Use a normal case and a busier case. Do not extrapolate exclusively from a quiet week or an unusually enthusiastic first day.

4. Choose continuity or a firm cutoff. To configure the meter, select Update extra usage limit and save the chosen value. To prevent extra usage, select Disable extra usage from its dropdown. Atlassian documents these controls under both Billing and Insights. Source: Manage extra usage for a meter.

5. Document what happens at the limit. Atlassian says new credit-consuming requests pause when the applicable allowance or spending cap is exhausted; ongoing Chat conversations and Agent tasks finish. Explain the fallback process to the team before a deadline depends on it. Source: Rovo licensing.

For a concrete planning threshold, suppose you allow $50 monthly overages and reserve 1,000 credits for other Rovo activity:

Credits available for MCP
= 500 included + ($50 ÷ $0.01) − 1,000 other
= 4,500 credits

That supports 450 prompts at ten credits each, or 180 at 25 credits each. Across ten people and 20 days, that is 2.25 or 0.9 prompts per person daily. These are calculated capacities, not individual allocations or guarantees about an in-flight cutoff.

Start with a spending limit tied to the workflows your team values, then replace assumptions with usage evidence. Saving a few dollars on predictable lookups is less consequential than discovering that your Friday handoff routinely needs several enriched calls.

Freshness checkpoint: Recheck this forecast on December 3, 2026, replacing future-billing language and reconfirming rates, allowances, defaults, and cutoff behavior. Refresh sooner if metered operations or per-call rules change. Recheck tool names for March 1, 2027, when Atlassian documents an automatic transition from existing MCP v1 usage to v2 tools. Source: Atlassian’s MCP getting-started guide.