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

# Usage & Limits

> How usage is counted and what limits apply to the NBQ Engine API.

## How usage is counted

Usage is counted **per conversation**, keyed by your `session_id`.

* The first call with a new `session_id` registers the conversation.
* Subsequent calls with the same `session_id` belong to that conversation and are not counted again.
* Your current consumption, remaining allowance, and any overage are visible on the **Dashboard**.

## Plan limits

Plans may include a number of conversations per period (monthly or annual). Depending on your contract, overage is either:

* **Billed** — conversations above your allowance are charged at the per-conversation rate.
* **Blocked** — requests return `HTTP 429` when the quota is reached.

## API field limits

| Field                            | Limit                         |
| -------------------------------- | ----------------------------- |
| `session_id`                     | 1–128 characters              |
| `conversation_history`           | Up to 50 messages             |
| `conversation_history[].content` | 1–2000 characters per message |
| `context`                        | Up to 8000 characters         |
| `answered_question_ids`          | Up to 200 entries             |
| Conversion `metadata`            | Serialised size ≤ 4096 bytes  |
| CSV import file size             | Up to 2 MB                    |

## Rate limiting

Rate limiting is enforced at the API level. When you exceed it, requests return **HTTP 429**. Back off and retry with exponential backoff.

## Minimum question bank

The API Key tab in the NBQ Studio is **locked** until your NBQ instance has at least **10 active questions**. The API cannot serve traffic until this threshold is met.