• MODEL LIST
  • SPEC DETAILS
  • CHAT RUN
  • HUMAN

How to route prompts through different LLM models with your AI

Connect Glama once to give your agent access to a wide range of models through a single gateway. You can then swap models based on the specific needs of your task.

Ask AI about this page

Short answer

How can I switch between different LLM models?

You use the Glama gateway to access a variety of models through one interface. Your agent can check available models and their specific details to decide which one fits your current prompt's context or budget requirements.

Where every request lands.

Model routing outcomes

The gateway handles the heavy lifting of model selection and execution.

MODEL LIST

View available models

The agent identifies which models are currently active on the gateway. You get a clear list of what's available to use.

SPEC DETAILS

Check model constraints

The agent looks up specific details like context window sizes. This helps you pick a model that won't cut off your text.

CHAT RUN

Execute the prompt

The agent sends your request to the chosen model via the gateway. You receive the direct response from the selected model.

HUMAN

Manual model selection

If the agent can't decide which model is best for your specific cost or context needs, you step in to pick the model yourself.

The workflow

What your AI does when a routing request arrives.

The agent uses the gateway to inspect and interact with the model ecosystem.

  1. Inventory models

    The agent pulls a list of all models currently accessible through the gateway to see what's on the table.

    glama_get_gateway_models

  2. Inspect model specs

    The agent looks at the specific attributes of a model, such as its context window, to ensure it can handle your prompt.

    glama_get_gateway_model_details

  3. Run the chat

    The agent sends the actual prompt to the specific model you've targeted through the gateway.

    glama_run_gateway_chat

  4. Verify model availability

    The agent re-checks the model list if a specific model is unavailable or if you need to switch to a cheaper option.

    glama_get_gateway_models

Try it

Copy these to start.

Use these prompts to begin routing your requests.

Starting points

These are starting points. Swap out the specific model names or task descriptions to match your actual needs.

Glama Connector

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

Connector linkhttps://edge.vinkius.com/vk_preview_3I3xkcQ0Lir2EbUp0HqzccT0Pcn6Iy0qjp9A5LCu/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 — Glama capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "glama-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_3I3xkcQ0Lir2EbUp0HqzccT0Pcn6Iy0qjp9A5LCu/mcp"
    }
  }
}
Copy into chat04

  • List all the models available through the Glama gateway.

  • Which model in the gateway has the largest context window?

  • Run a chat prompt using the model with the lowest cost.

  • Check the details for the model named 'gpt-4o' in the gateway.

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

Start here

Connect Glama once, then ask.

Just link your account once. Your credentials stay encrypted, and you can immediately start asking your agent to switch models.

Connect Glama to your AI

FAQ

How the gateway behaves

  • 01

    Can the AI automatically switch models based on my budget?

    Yes, if you tell the agent to find the most cost-effective model using the gateway tools, it can do that.

  • 02

    Can the AI change my Glama account settings?

    No. The tools only allow the agent to read model info and run chats. It cannot modify your account or billing.

  • 03

    Does the AI decide which model to use every time?

    It can, but you can also explicitly tell it which model to use for a specific prompt.

  • 04

    Can the AI delete my chat history in Glama?

    No. There are no tools provided to delete or modify existing chat history.

  • 05

    How does the AI know if a model has a large enough context window?

    The agent uses the tool to look up the specific details of a model to check its limits before sending the prompt.

  • More questions about Glama? See everything the Glama Connector can do

Connect Glama to Claude, Cursor, ChatGPT & more