- STATUS
- RESOURCES
- CONFIG
- HUMAN
Short answer
How can I audit my model deployments using AI?
Your AI client pulls real-time data from Baseten to check replica counts, autoscaling settings, and inference bounds. You get a clear view of your active infrastructure through direct queries. This breakdown shows how your agent handles the audit.
Audit outcomes
Where your data lands.
What happens when your agent scans your Baseten setup.
STATUS
Live deployment checks
The AI pulls current replica counts and scaling configurations. You see exactly how your models are running right now.
RESOURCES
Model inventory
Your agent lists every model in your account. This helps you spot unused or redundant models quickly.
CONFIG
Scaling audits
The AI reviews autoscaling limits and inference bounds. You can identify if your settings match your actual traffic.
HUMAN
Manual adjustments
The AI identifies issues but cannot change your settings. You must log in to Baseten to fix any misconfigurations.
The workflow
What your AI does when the audit request arrives.
The agent moves from broad discovery to specific configuration checks.
Inventory scan
Your agent starts by pulling a full list of all models available in your Baseten account.
list_modelsDeployment discovery
The AI identifies every active deployment currently running those models.
list_deploymentsDeep dive
Your agent pulls specific configuration details for a single deployment to check its bounds.
get_deploymentReport generation
The AI compares the retrieved replica counts and scaling settings against your requirements.
get_deployment
Try it
Copy these to start.
Use these prompts to begin your audit.
Starting points
These are starting points. Swap out the specific deployment names or model IDs to match your actual Baseten setup.
Baseten Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_Iok4x74580t8DfXgX0kQJnXTan6vttePrqZ5XRGU/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Baseten capabilities are ready to use.
{
"mcpServers": {
"baseten-mcp": {
"url": "https://edge.vinkius.com/vk_preview_Iok4x74580t8DfXgX0kQJnXTan6vttePrqZ5XRGU/mcp"
}
}
}List all my models in Baseten and tell me which ones are currently deployed.
Check the replica count and autoscaling settings for my 'llama-3-70b' deployment.
Are any of my Baseten deployments running with insufficient replica counts?
Show me the inference bounds for the 'stable-diffusion-v1' model deployment.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Baseten once, then ask.
You only need to link your account once. Your credentials stay encrypted, and you can immediately start asking questions about your deployments.
Connect Baseten to your AIFAQ
How this task behaves
- 01
Can the AI change my autoscaling settings?
No. The AI can only read your current configurations. You have to make any changes manually in the Baseten dashboard.
- 02
Can the AI delete a deployment if it looks inefficient?
No. The AI has no permission to delete or modify your deployments.
- 03
How does the AI know which models I have?
It uses the list_models tool to fetch your entire model inventory.
- 04
Can I audit my deployment costs?
No. This connector only provides access to deployment counts, scaling, and model data.
- 05
Does the AI see my model weights?
No. The AI only sees the deployment metadata and configuration settings.
More questions about Baseten? The Connector page answers them. See everything the Baseten Connector can do
Connect Baseten to Claude, Cursor, ChatGPT & more
