LLM Integrations

LLM (Large Language Model) Integrations connect Morpheus to language model providers. These integrations supply the AI reasoning capabilities that power AI Agents.

Navigate to Tools > AI Services > Integrations to manage connected LLM providers.

Adding an LLM Integration

LLM integrations are managed from Tools > AI Services > Integrations:

  1. Navigate to Tools > AI Services > Integrations

  2. Click + New Integration

  3. From the New Integration dropdown, select a type

    Note

    Currently, all integration types consume the same configuration fields.

  4. Configure the following fields:

    • NAME — A descriptive name for this integration

    • ENABLED — Check to activate the integration

    • API ENDPOINT — The URL for the provider’s API endpoint

    • CREDENTIALS — Select Local Credentials to paste an API key directly, or select from the secure credential store (e.g., Cypher)

  5. Save the integration

After saving, Morpheus automatically discovers available models from the provider and syncs them. These models then appear in the Model dropdown when configuring AI Agents.

Supported Providers

Morpheus ships with provider integrations pre-installed but more will be added over time. Additionally, new AI integrations can be added to the platform by users through custom plugin development.

The following providers are currently available:

Provider

Notes

GitHub Copilot

GitHub Copilot integration. Requires GitHub Copilot subscription and authentication token.

Ollama (Local)

Self-hosted open-source models (Llama, Mistral, DeepSeek, etc.). Requires Ollama server URL. No API key needed for local instances.

OpenAI-Compatible (Local)

Any local LLM server that exposes an OpenAI-compatible API (e.g., LM Studio, vLLM, LocalAI). Configure with the local server endpoint URL.

Note

Additional LLM providers can be added through the Morpheus plugin system. See the developer documentation for guidance on creating custom AI integration plugins.

Model Properties

Discovered models include metadata to help select the appropriate model for your use case:

Property

Description

Context Window

Maximum number of tokens the model can process in a single conversation (input + output)

Max Output Tokens

Maximum tokens the model can generate in a single response

Speed Score

Relative speed rating for response generation

Quality Score

Relative quality/capability rating

Cost Score

Relative cost per request/token

Usage Tracking

LLM integrations track token and request usage:

  • Token Usage — Total tokens consumed, remaining quota, and reset period

  • Request Usage — Total requests made, remaining quota, and reset period

These metrics help monitor API consumption and manage costs. Usage data is visible on the integration detail page.

Choosing a Model

When selecting a model for an AI Agent, consider:

  • Context window — Larger windows allow longer conversations and more tool results. Models with 128K+ context windows are recommended for complex infrastructure queries.

  • Tool calling support — Ensure the model supports function/tool calling (most modern models do). This is required for MCP tool invocation.

  • Speed vs. quality — Faster models (GPT-4o-mini, Claude 3 Haiku) are better for frequent, simple queries. Higher-quality models (GPT-4o, Claude 3 Opus) are better for complex reasoning and multi-step operations.

  • Cost — High-throughput environments with many users should consider cost per token carefully.