Skip to main content

Ollama

Ollama is a local-only model runner. MCP support depends on the client you pair it with — the Ollama model itself doesn't speak MCP, but agent wrappers like LangGraph, Continue.dev, or a custom script can bridge Ollama to Versionx's MCP server.

Two ways to use Ollama with Versionx​

1. BYO-API-key mode (no MCP)​

Point the --ai-* flags at your local Ollama endpoint:

export OLLAMA_BASE_URL=http://127.0.0.1:11434
export VERSIONX_AI_PROVIDER=ollama
versionx release plan --ai-changelog

This uses Ollama for one-shot LLM calls without any MCP dance. Simplest path for local-only use.

2. As the driver for MCP (via an agent wrapper)​

Any MCP-aware agent that supports configurable LLM backends can combine Ollama and Versionx. Examples:

  • Continue.dev. Configure Versionx as an MCP server and point Continue's model settings at Ollama. Tool-calling-capable Ollama models (e.g., qwen2.5-coder, llama3.3) work best.
  • Custom script. The Versionx MCP server runs as a subprocess; your script reads Ollama-generated tool calls and forwards them. The rmcp SDK on PyPI and npm can handle the client side.

Local + private​

Neither path sends data anywhere beyond your machine. Ollama is local; Versionx never phones home; the MCP server is stdio-only by default.

See also​