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

# FAQ

> Common questions about using Metrion.

<AccordionGroup>
  <Accordion title="Does Metrion add latency to my requests?">
    Yes, a small amount. Metrion is hosted in EU West (Ireland), so there is an additional network hop between your application and the AI provider. In practice, the overhead is typically a few milliseconds and is negligible compared to the provider's own response time.

    You can see the actual end-to-end latency for every request on the **Logs** page and in the **Latency** tab of your dashboard.
  </Accordion>

  <Accordion title="Is my provider API key secure?">
    Yes. In stored mode, your key is encrypted at rest. The encryption key is stored separately from the database, so your key material is never accessible in plain text. Metrion never logs or exposes key values anywhere in the application.

    In pass-through mode, your key is forwarded directly to the provider on each request and is never stored by Metrion.
  </Accordion>

  <Accordion title="Does Metrion work with streaming?">
    Yes. Streaming (`stream: true`) works for all supported providers. Pass it in your request body exactly as you would with the provider's API directly — no additional configuration is needed on your end.
  </Accordion>

  <Accordion title="What happens if I hit the Free plan limit?">
    After 10,000 requests in a calendar month, the proxy returns a `429 Too Many Requests` response for any further requests. Your existing data — logs, dashboard stats, and usage history — remains fully accessible.

    The counter resets automatically on the first of each month. To remove the monthly cap entirely, upgrade to Pro.
  </Accordion>

  <Accordion title="Can I use Metrion without storing my provider keys?">
    Yes. Use pass-through mode: pass your provider API key directly as the `api_key` (for Anthropic) or `Authorization: Bearer` token (for OpenAI-compatible providers). Metrion detects that it is not a Metrion token and forwards it to the provider.

    Your usage is still tracked and visible in your dashboard. Your key is never stored.
  </Accordion>

  <Accordion title="What providers are supported?">
    Metrion currently supports:

    * Anthropic
    * OpenAI
    * Google Gemini
    * Mistral
    * Grok (xAI)

    More providers are coming. Check the [Integrate page](https://www.metrion.dev/integrate) for the latest list.
  </Accordion>

  <Accordion title="How do I track usage per user or bot?">
    Add the `x-metrion-user` header to your requests with a string identifier — a user ID, an agent name, or a workflow label. This value appears in the **Requester** column in your logs and can be filtered from the dashboard.

    See [Request Headers](/reference/headers) for code examples.
  </Accordion>

  <Accordion title="How do I export my data?">
    * **Logs export (Pro):** On the Logs page, click **Export CSV** to download your request logs.
    * **Full account export:** Go to **Settings → Profile → Download my data** to download a complete JSON export of your account, including alert rules, usage records, and logs.
  </Accordion>

  <Accordion title="Does Metrion modify my requests or responses?">
    No. Metrion forwards your request body to the provider as-is and returns the provider's response unchanged. The only exception is the `anthropic-version` header, which Metrion sets to a fixed value server-side for Anthropic requests.

    Metrion reads usage metadata (token counts, latency) from the response to populate your dashboard, but does not alter the response content.
  </Accordion>

  <Accordion title="Where is my data stored?">
    All data is stored in the EU West region (Ireland) for GDPR compliance.
  </Accordion>
</AccordionGroup>