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

# View Logs

You can view your recent call logs in the console: [https://poixe.com/log](https://poixe.com/log)

Logs help you verify usage and costs, and troubleshoot request latency and routing. The platform records every request that was **successfully billed**, and retains only the **most recent 3 days** of data (on a rolling basis).

## What's in the logs

Each log entry typically shows the following information (fields may vary slightly across models/protocols):

* **Time**: When the request occurred
* **Token name**: The API token used for the request
* **Requested model / Processed model**: The model name you requested, and the model identifier the system actually processed
* **Requested vendor / Processed vendor**: Whether the default routing was used, and which vendor ultimately handled the request
* **Token counts**: Prompt, completion, and cache (if any)
* **Cost information**: Theoretical consumption, actual spend, and discount factor
* **Latency**: Upstream latency, system latency, and total latency
* **Response type**: For example, streaming / non-streaming

## How to find a log entry

On the logs page you can filter and search by common fields, for example:

* Filter by **token name** to find the request records of a given application
* Filter by **model name** to check calls to a particular model
* Use keywords to quickly locate an abnormal request (for example, long latency or unusual cost)

## View log details

Click **Details** on a log entry to see a fuller request summary, for example:

* Request ID, creation time, and response status
* Routing information (default routing / specified vendor, and the vendor that ultimately handled it)
* More complete token and billing details
* A more precise latency breakdown (system / upstream / total)
* Request origin information (such as IP)

This information is usually enough for:

* Reconciliation: Verify whether the charge for a call matches your expectations
* Performance troubleshooting: Determine whether latency mainly occurs upstream or on the network/system side
* Routing confirmation: Confirm whether the default routing was used and who actually handled the request

## Retention policy

* Logs are retained for the **most recent 3 days** only
* A rolling policy applies: new records are continuously written, and expired data is removed automatically
* If you need a longer period for auditing or reconciliation, we recommend saving key fields on your own side (for example, request ID, model, cost, and time)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.