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

# Agent Intelligence

> Choose models and manage spend for your agents

When you create an agent in AgentVault, you configure its intelligence settings:

* **Primary model** — the LLM your agent uses by default
* **Fallback chain** — alternates tried if the primary is unavailable or rate-limited
* **Daily and monthly spend caps** — hard limits on how much this agent can spend

## Setting intelligence during agent creation

The Agent Builder includes an **Intelligence** stage where you configure these settings before deploying. Defaults are pre-filled and work out of the box.

## Changing settings after creation

Open your agent's detail page → tap the gear icon (top right) → **Intelligence** tab. Edit and save. Changes take effect on the next message.

## Spend caps

Caps are enforced by the LiteLLM proxy before a request reaches the model. When an agent hits its cap:

1. The agent stops responding for that period
2. A banner appears in the conversation with a **Raise cap** button (doubles the current cap)
3. You can also open **Agent settings → Intelligence** to set a precise new cap

Caps reset on schedule:

* **Daily** — midnight UTC
* **Monthly** — first of the month UTC

## Available models

| Model                         | Best for                             |
| ----------------------------- | ------------------------------------ |
| `anthropic/claude-sonnet-4-6` | Complex reasoning, coordinator tasks |
| `anthropic/claude-haiku-4-5`  | Fast replies, brainstorming          |
| `openai/gpt-4o`               | General use                          |
| `openai/gpt-4o-mini`          | Cost-sensitive, high-volume          |

Your platform administrator can add additional providers via the ops runbook.