Does it provide performance metrics?

Yes, it shows how to extract prompt evaluation counts, evaluation duration, and calculate tokens per second from the API response.

What is the advantage of using this skill over a high-level library?

It provides direct control over HTTP requests, allowing for easier debugging, custom headers, and access to raw response metrics that high-level abstractions might hide.

What is the default connection endpoint?

It defaults to http://localhost:11434 but is fully configurable via the OLLAMA_HOST environment variable for containerized or remote environments.

Can I manage my local models with this tool?

Absolutely. The skill covers endpoints for listing available models, viewing specific metadata, copying models, and deleting them from your local instance.

Does this skill support streaming responses?

Yes, it includes specific implementation patterns for handling streaming text generation and chat responses using iter_lines() for real-time output.

Ollama REST API Interface

Name: Ollama REST API Interface
Author: atrawog

byatrawog

Data Science & ML

Integrates Ollama's local LLM capabilities into Claude Code via direct REST API operations using the Python requests library.

About

This skill empowers Claude to interact directly with locally hosted Ollama instances through its native REST API. It provides comprehensive control over model lifecycles—including listing, pulling, and deleting models—as well as executing text generation, chat completions, and embedding tasks. By using the standard requests library instead of high-level wrappers, it offers developers a low-level, highly customizable way to build private AI workflows, debug API interactions, and handle streaming responses with full access to response metrics and token usage data.

Key Features

Detailed performance tracking including token-per-second and eval metrics
Local embedding generation for RAG and vector search workflows
Support for both streaming and non-streaming response patterns
0 GitHub stars
Complete coverage of Ollama /api/* endpoints for chat and generation
Comprehensive model management including list, show, copy, and delete

Use Cases

Managing local model inventory and VRAM usage within the terminal
Building custom local AI pipelines and automated workflows
Debugging and monitoring raw Ollama API responses and performance

About

Key Features

Detailed performance tracking including token-per-second and eval metrics
Local embedding generation for RAG and vector search workflows
Support for both streaming and non-streaming response patterns
0 GitHub stars
Complete coverage of Ollama /api/* endpoints for chat and generation
Comprehensive model management including list, show, copy, and delete

Use Cases

Managing local model inventory and VRAM usage within the terminal
Building custom local AI pipelines and automated workflows
Debugging and monitoring raw Ollama API responses and performance