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

# Billing Agent

# Billing

Condense AI Agents use a token-based billing model that provides predictable consumption tracking and transparent cost management. Every AI interaction consumes input and output tokens, which are tracked independently and deducted from environment-level token wallets.

The billing system is designed to support both enterprise contracts and self-service organizations while providing complete visibility into AI usage across users, workspaces, environments, and agents.

## How Billing Works

Each Condense Environment (ENV) is provisioned with two independent token wallets:

* Input Token Wallet
* Output Token Wallet

Whenever an AI Agent processes a request:

1. The platform estimates the required token usage.
2. Available wallet balance is verified.
3. The request is executed if sufficient balance exists.
4. Actual token consumption is recorded.
5. Input and output wallets are updated.
6. Usage is logged for reporting and billing.

All token usage is tracked automatically and is available through the Agent Usage dashboard.

## Billing Model

Condense supports prepaid token-based billing.

Organizations receive token allocations based on their subscription or contract. Additional tokens can be purchased whenever required.

Billing supports:

* Monthly token allocations
* Manual token purchases
* Contract-specific pricing
* Master rate card pricing
* Usage reporting
* Transaction history

## Token Consumption

Every AI request consumes two categories of tokens.

| Token Type    | Description                                                            |
| ------------- | ---------------------------------------------------------------------- |
| Input Tokens  | Tokens sent to the LLM as part of the prompt and conversation context. |
| Output Tokens | Tokens generated by the LLM in the response.                           |

Both token types are tracked independently because different AI providers may charge different rates for input and output processing.

## Usage Tracking

Every AI request is recorded for billing and auditing purposes.

Tracked information includes:

* User
* Environment
* Workspace
* AI Agent
* LLM Model
* Prompt
* Timestamp
* Input Tokens
* Output Tokens

This information is available through the Agent Usage dashboard and transaction history.

## Recharge Tokens

Environment administrators can purchase additional tokens whenever required.

Purchased tokens are immediately credited to the environment wallets and become available for all AI Agents running within that environment.

## Access Control

Billing operations are protected through role-based access control (RBAC).

Different permissions apply to:

* Viewing token balances
* Viewing usage history
* Purchasing additional tokens
* Managing billing settings

Only authorized administrators can purchase additional tokens for an environment.

## Related Pages

* Token Wallets
* Agent Usage
* Recharge Tokens
* Pricing & Billing