- DISCOVERY
- EXECUTION
- STRUCTURE
- HUMAN
Short answer
How can I run NLP inference with my AI?
Your agent uses Metatext to find specific models and send text to them for immediate predictions. You get the results directly in your chat interface. This setup gives you access to structured model outputs for any text you process.
The outcomes.
What you get from the AI.
The agent manages the connection between your text and the deployed models.
DISCOVERY
Model selection
The agent identifies which models in the Metatext library can handle your specific text task.
EXECUTION
Real-time predictions
The agent sends your text to a model and returns the prediction results immediately.
STRUCTURE
Parsed data
The agent turns raw model outputs into clean, readable information like labels or scores.
HUMAN
Final verification
You review the model's predictions to ensure they are accurate before you move forward with the data.
The workflow
What your AI does when text arrives.
Your agent handles the technical steps of interacting with the Metatext API.
Find a model
The agent scans the Metatext catalog to find a model that fits your specific request.
search_nlp_modelsCheck requirements
The agent looks up the specific input needs for the chosen model to ensure the data is compatible.
get_model_detailsRun inference
The agent sends the text to the deployed model to generate a prediction.
run_model_inferenceDeliver results
The agent takes the model output and presents the final prediction to you in the chat.
run_model_inference
Try it
Copy these to start.
Use these prompts to trigger the Metatext tools.
Starting points
Replace the bracketed text with your specific text or model names.
Metatext Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_77VAYStySeROKtVdW5OxQRJpWkfEWaDXQ23SvhBu/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 — Metatext capabilities are ready to use.
{
"mcpServers": {
"metatext-mcp": {
"url": "https://edge.vinkius.com/vk_preview_77VAYStySeROKtVdW5OxQRJpWkfEWaDXQ23SvhBu/mcp"
}
}
}Search for NLP models that can perform sentiment analysis on this customer feedback: [paste text here]
Find the details for the model named [model_name] so we can check its input requirements.
Run inference on this paragraph using the [model_name] model: [paste text here]
Check which models are available for entity extraction in the Metatext library.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Metatext once, then ask.
Click the link below to connect Metatext to your AI client. Your credentials stay encrypted on our side.
Connect Metatext to your AIFAQ
How this task behaves.
- 01
Can the AI create new NLP models?
No. The AI can only search for, inspect, and run inference on models that are already deployed in Metatext.
- 02
Can the AI change model settings?
No. The AI uses existing model configurations; it cannot modify the underlying model parameters.
- 03
Can I use this with Cursor or Claude?
Yes. Any MCP-compatible client can use these tools once you connect the Metatext MCP.
- 04
Will the AI run models on everything I type?
No. The AI only runs inference when you specifically ask it to use the Metatext tools.
- 05
Can the AI delete models from my account?
No. There are no tools provided to delete or remove models from Metatext.
More questions about Metatext? The Connector page answers them. See everything the Metatext Connector can do
Connect Metatext to Claude, Cursor, ChatGPT & more
