> ## Documentation Index
> Fetch the complete documentation index at: https://www.vollna.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Changes and limits

> Which changes ask you first, and what each call uses from your plan

## Changes

Most changes run when the app calls them, the same as a click in the dashboard. Your AI app may also ask you before each call. Each app has a setting for this per tool.

## Changes that ask you first

Vollna holds back the changes that spend connects or money, reach other people or cannot be undone. The tool returns a preview first and nothing changes until you agree. The app then sends the change again with your confirmation.

| Change | The preview shows |
| - | - |
| A bid: schedule, approve or release a job, or bid on a job the qualifier rejected | The connects it costs, the agency balance and the limits left |
| Delete a filter | The auto-bidding setups that pause |
| Delete a template | The filters that stop offering it and the finished A/B tests it leaves |
| Activate an auto-bidding setup | Its filter and template, the daily proposal limit, the boost ceiling and the agency's connects balance |
| Pause an auto-bidding setup | The queued jobs it skips |
| Delete an auto-bidding setup | That it stops bidding, and its jobs stay in the history |
| End an A/B test | The template that stays and the ones dropped |
| Buy connects now | Up to how many connects, the payments, this month's top-ups and the agency's balance |

A preview is valid for 10 minutes and works once. When the data changes in the meantime, the app gets a new preview to show you.

These changes ask first only in some cases:

| Change | Asks first when |
| - | - |
| Save a qualifier prompt | Other filters or auto-bidding setups share the qualifier |
| Change a proposal template | An active auto-bidding setup sends with it |
| Change an auto-bidding setup | The setup is active and the change raises spending, rewrites a shared qualifier or restarts an A/B test |
| Change auto-bidding settings | It turns auto top-up on or raises its limits |
| Over-limit billing | It turns over-limit billing on |

[Tools](/docs/mcp/tools) marks each tool as Reads, Changes or Asks first.

## Plan limits

MCP has the same plan limits as the dashboard: filters, active filters, AI qualifier credits, proposal credits and auto-bidding limits. A call past a limit is refused with the limit and when it resets.

## Credits

| What the app does | What it uses |
| - | - |
| Reads jobs: feed jobs, verdicts, hidden jobs, preview samples, auto-bidding jobs | 1 project credit per job, the first time it reads that job in a usage period. The jobs also count toward your feed's daily job cap |
| Qualifies jobs, or tests a qualifier prompt on past jobs | AI qualifier credits, the same pool as the dashboard. The answer says how many |
| Writes a test proposal | The AI proposal preview quota of the dashboard |

The [API](/docs/api-reference/introduction#credit-system) uses the same project credits.

## Daily caps

Some reads have a daily cap per team. MCP and the API share these caps. They reset at 00:00 UTC.

| Read | Default per day |
| - | - |
| Why a job is or is not in a filter | 100 |
| Test a filter change on the last 30 days | 100 |
| Keyword suggestions for a filter | 20 |
| Market Insights | 50 |
| Niche difficulty, competition on a job | 50 |
| Keyword opportunities | 20 |
| Reply rate diagnosis | 20 |
| Draft template instructions | 10 |

Your plan can set other values. Ask the app "How much do I have left today?", or open [Settings → Developers → MCP](https://www.vollna.com/dashboard/team/mcp).

You can make up to 60 calls a minute, across all your connected apps.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.