Ollama Handoff

Offloads cheap work from an AI agent to a local Ollama model, keeping that work off paid cloud models.

What you can do

• Hand off lightweight subtasks from the assistant to a local Ollama model instead of a cloud model.
• Choose the default model, context window, request timeout and how long the model stays loaded in VRAM through environment variables.
• Point the handoff at a specific Ollama server base URL.
• No tool list is published, so the exact operations are not documented in the supplied material.

When to use it

Useful when you already run Ollama locally and want an agent to push simple, low-stakes work to a local model to avoid cloud token costs, while keeping heavier reasoning on the main model.

Requirements

Runs as a local process on the user's machine, installed from the PyPI package ollama-handoff over stdio. Needs a reachable Ollama server, configured with OLLAMA_URL, and optionally OLLAMA_DEFAULT_MODEL, OLLAMA_KEEP_ALIVE, OLLAMA_NUM_CTX and OLLAMA_TIMEOUT_S. Desktop only; no web execution. No authentication is declared.