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

# Managing Costs

> Pricing model, balance management, and auto-deposit

## How Pricing Works

* Each dataset has a **price per chunk** (e.g., `$0.001`)
* A small **platform fee** is added per chunk
* You're charged for each result **actually returned**, not the `top_k` estimate
* Cost = `chunk_count × (price_per_chunk + platform_fee)` per dataset

### Cost Estimation

Before a query executes, Datagate checks your balance against the **max possible cost** (topK × price per dataset). If your balance is too low, you get a `402 Insufficient Balance` error with your current balance and the estimated cost.

## Deposits

Add funds via Stripe Checkout from the [dashboard](https://platform.getdatagate.com):

1. Go to **Billing** → **Deposit**
2. Enter amount and complete Stripe Checkout
3. Funds are credited immediately after payment

## Auto-Deposit

Configure automatic refills so your agent never runs out of balance:

1. Go to **Billing** → **Auto-Deposit** in the dashboard
2. Set a **threshold** (e.g., \$5.00) — when balance drops below this, auto-refill triggers
3. Set a **refill amount** (e.g., \$25.00) — how much to deposit
4. Requires a saved payment method (set up during your first deposit)

Auto-deposit triggers after each query if your balance dropped below the threshold. It uses your saved card with off-session payment consent.

## Monitoring Balance

Check your balance from the dashboard or programmatically via the API:

```bash theme={null}
curl -H "Authorization: Bearer dg_live_..." \
  https://api.getdatagate.com/v1/billing/buyer-balance
```

Response:

```json theme={null}
{
  "balance": "24.50000000",
  "total_deposited": "100.00000000",
  "total_spent": "75.50000000",
  "currency": "USD"
}
```

## Tips for Cost Management

* Use **lower `top_k`** values when you don't need many results (default is 10)
* Use **metadata filters** to narrow results — fewer irrelevant chunks = lower cost
* Set up **auto-deposit** with a comfortable threshold for unattended agents
* Check `price_per_chunk` in `list_datasets()` to understand per-dataset pricing before querying