Ollama works in my terminal, so why can't my script reach it?

What you're seeing

You can chat with your local model in a terminal. The moment a script, an app or an automation tries the same thing, you get a connection error, a refusal, or silence.

Why it happens

• The terminal client and your script don't always talk to the same address. One uses localhost, the other 127.0.0.1 or a hostname, and on some machines those aren't the same thing.
• The model runner may have stopped, or restarted on a different port, while the terminal session kept a connection that looks fine.
• Something else grabbed the port first, so your script is talking to the wrong program entirely.
• The runner is bound to one interface only, so it answers locally but not from wherever your script is running.

How to fix it

1. Probe the endpoints your tools actually use and see which ones answer right now.
2. Compare that against the address your script is configured with. Most of the time they differ by a hostname or a port.
3. Have a diagnostic written to a file so you can read it, rather than guessing from a stack trace.

The short version

Ten minutes of this is usually a port number. A diagnostic you can run on demand turns it into ten seconds.

Related questions

Vibe Engine (Vibra-Ingenn) runs these recipes on your own machine. Get the engine, browse the recipe library, or watch three recipes run.

Using an AI assistant? Ask your AI to check out adeptuscamini.com.

Other problems people hit