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

# Plans and usage

> What each plan allows in agents, test conversations a month, own rules and report review, and what happens at the limit.

Your plan is counted in test conversations a month. One test conversation is one run of a test: our attacker talks to one agent over one channel until it succeeds or gives up.

## The plans

| Plan | Agents | Test conversations a month | Own rules | Reports | Separate attacks per goal |
| - | - | - | - | - | - |
| Essentials | 1 | 25 | No, standard rules only | Automatic, not reviewed | 16 |
| Pro | 3 | 75 | Yes | Automatic, not reviewed | 24 |
| Team | No limit | 500 | Yes | Reviewed by DeepBlocker | 40 |
| Business | No limit | 2,500 | Yes | Reviewed by DeepBlocker | 80 |
| Enterprise | No limit | 10,000 | Yes | Reviewed by DeepBlocker | 160 |
| Custom | Set by DeepBlocker | Set by DeepBlocker | Yes | Reviewed by DeepBlocker | Set by DeepBlocker |

On Essentials and Pro, each extra agent adds one agent and 25 test conversations. Prices are on [Pricing](https://docs.deepblocker.ai/account/pricing.md).

## Channels per plan

| Plan | Channels in the plan | Channels you can test on today |
| - | - | - |
| Essentials | Web or API, plus one chat channel | Web chat or API |
| Pro | Every chat channel | Web chat or API, Telegram |
| Team and above | Every channel | Web chat or API, Telegram |

Only the two channels in the last column run today. See [Channels](https://docs.deepblocker.ai/agent-assurance/channels.md).

## How a conversation is counted

| Rule | Detail |
| - | - |
| Month | The calendar month, in UTC |
| Counted when | A run starts. Two runs starting together cannot both take the last one |
| Given back when | The run never started, for example because the sandbox could not open it |
| A test that repeats | Uses one conversation each time it runs |
| A sweep | Uses its whole budget, shown on the form before you save |
| Messages per conversation | Short 6, Standard 10, Long 16. No test ever passes 30 messages |

## Usage states

The **Overview** tab shows "N of M test conversations" with a bar.

| State | When | What you see |
| - | - | - |
| OK | Under 80 percent used | A blue bar |
| Near limit | 80 percent or more | An amber bar from 75 percent, red from 90 percent |
| Used up | 100 percent | A red bar, and no new test can start |

## What happens at the limit

| Limit | What happens |
| - | - |
| Agents | **Add an agent** is greyed out with "Your plan allows N agents and all are in use. Remove one, or ask to add agents to your plan." |
| Conversations, when planning | The form says "That is more than you have left. Make the test smaller, or wait for next month." |
| Conversations, while running | A test that needs the next conversation pauses with the reason "This workspace has used its N test conversations for the month". It resumes when the month turns or the plan grows |
| A channel removed from the plan | A test on that channel pauses |

## Where usage is shown

* The **Overview** tab: plan name, conversations, agents, renewal date and the next test.
* The agent's **Tests** tab: "N of M test conversations used this month, N left."
* The plan form: "This test uses N test conversations. You have N left this month."

See also [Usage](https://docs.deepblocker.ai/account/usage.md) for your other DeepBlocker licences.


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