• PREDICTION
  • DATA FLOW
  • HUMAN
  • INTEGRATION

How to run model inference with your AI client

Connect your deployed models to your AI client. Once linked, your agent can pass data to your models and return results directly in your conversation.

Ask AI about this page

Short answer

How can I run model inference with my AI?

Your agent uses the get_inference tool to send data to your deployed models. You get the prediction results immediately in your chat. This setup lets you use custom ML models without leaving your conversation.

The outcomes

Where your data goes

The AI handles the communication between your text and your models.

PREDICTION

Instant model results

The agent sends your input to the model. You get the prediction result right in the chat window.

DATA FLOW

Automated data passing

The AI extracts the necessary values from your prompt. This means you don't have to manually format JSON payloads.

HUMAN

Manual review needed

The AI shows you the model's output. You decide if the prediction is correct or if you need to try a different input.

INTEGRATION

Live model access

Your agent connects to your deployed endpoint. This turns your chat into a functional interface for your ML models.

The workflow

What your AI does when you request inference.

The AI acts as the bridge between your text and your model endpoint.

  1. Identify data

    Your agent reads your prompt to find the specific parameters required by your model.

    get_inference

  2. Call the model

    The AI sends the extracted data to your deployed endpoint to trigger the prediction.

    get_inference

  3. Process response

    The agent receives the raw output from the model and prepares it for reading.

    get_inference

  4. Display result

    The prediction appears in your chat window so you can act on it.

    get_inference

Try it

Copy these to start.

Use these prompts to trigger your first model prediction.

Starting points

Swap out the bracketed text for your actual model names and data.

Modelbit Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_U02S7R70GaoTP6dztOIIMDtmgDyuN2GTH3Or66r0/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Modelbit capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "modelbit-ml-model-deployments-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_U02S7R70GaoTP6dztOIIMDtmgDyuN2GTH3Or66r0/mcp"
    }
  }
}
Copy into chat04

  • Run inference on my 'sentiment-model' using this text: [insert text]

  • Use the 'image-classifier' model to predict this description: [insert description]

  • Send this customer feedback to my 'toxicity-detector' model and show me the result.

  • Run the 'price-predictor' model on these features: [insert feature list]

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Modelbit once, then ask.

One link connects your models to your AI client. Your credentials stay encrypted on our side.

Connect Modelbit to your AI

FAQ

How this task behaves

  • 01

    Can the AI deploy new models for me?

    No. This MCP only runs inference on models you have already deployed.

  • 02

    Can the AI change my model's configuration?

    No. The AI only sends data to the model and receives the output. It cannot modify your model settings.

  • 03

    Does the AI automatically take actions based on predictions?

    No. The AI provides the prediction to you. You must decide if you want to take further action.

  • 04

    Can I use any model via this MCP?

    You can use any model that is deployed through Modelbit and accessible via the get_inference tool.

  • 05

    Does the AI delete my model data?

    No. The AI only uses the data provided in your prompt to perform the inference.

  • More questions about Modelbit? The Connector page answers them. See everything the Modelbit Connector can do

Connect Modelbit to Claude, Cursor, ChatGPT & more