> ## Documentation Index
> Fetch the complete documentation index at: https://docs.runwita.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI provider not configured

> The error you see when an AI feature runs without a configured provider, and how to fix it.

If you click **Process with AI** on the Inbox, or **Analyse** / **Reanalyse** on a journey's intelligence panel, without first configuring an AI provider in Settings, Runwita stops you with a clear error rather than failing silently or producing garbage:

> *AI Insights: No API key configured for provider "claude". Open Settings → Models and pick a provider before running this.*

The wording varies by which feature you triggered (the Frontier tier surfaces *AI Insights*; the Workhorse tier surfaces *Other AI features*) and by which field is missing (API key, base URL, or model). The fix is the same.

## How to fix

<Steps>
  <Step title="Open Settings → Models">
    Two cards: Frontier and Workhorse. The error message tells you which tier is the problem; the *other* tier might already be configured fine.
  </Step>

  <Step title="Pick a provider for the affected tier">
    Claude, OpenAI, Ollama, or Custom. See [Models](/settings/models) for the provider-by-provider walkthrough.
  </Step>

  <Step title="Paste your API key (if cloud), set base URL (if Ollama or Custom), pick a model">
    The model dropdown populates from the provider's `/v1/models` endpoint when you click *Load models*.
  </Step>

  <Step title="Save all">
    Bottom save bar. Then return to whatever you were doing and try again.
  </Step>
</Steps>

## Why the error exists

Before this guard was added, running an AI feature without a configured provider would fail at the network layer: a 401 from the provider, a DNS error, or a generic *"fetch failed"*. None of those tell you the actual problem. The new error surfaces the missing config explicitly and points you at the right Settings panel.

## When you'd see this error legitimately

Every time you reinstall Runwita on a fresh machine, before you've set up Models. Once configured, the error never reappears unless you've actively cleared a key or switched to a provider you haven't set up.

## What's next

<Card title="Settings: models" icon="sliders" href="/settings/models">
  The full model configuration walkthrough.
</Card>