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

# Token Factory FAQ — Inference, Data Retention, and Support

> Answers about Token Factory models, open-weight model API data retention, contractual terms, rate limits, and support.

Find answers to common questions about Corvex Token Factory. For data handling,
read the [Data Security page](/security/data-security), including retained
operational metadata and volatile prompt caching.

## How do I get an API key and make my first call?

<Accordion title="Create a key and send a request" defaultOpen>
  Sign in to the [Token Factory dashboard](https://tokenfactory.corvex.cloud/app),
  create an API key, and follow the [Quickstart](/getting-started/quickstart).
  Token Factory is currently in a free Alpha; paid top-ups are not required.
  See [Pricing](/pricing) for current billing status.
</Accordion>

## Can I switch without rewriting my code?

<Accordion title="Configure an existing client" defaultOpen>
  For supported OpenAI SDK calls, set the base URL to
  `https://api.tokenfactory.corvex.cloud/v1` and use your Token Factory API key.
  Compatibility depends on the endpoint, fields, and model features your
  application uses. The [OpenAI SDK guide](/integrations/openai-drop-in) includes
  Python and TypeScript examples and compatibility boundaries.
</Accordion>

## What happens to my prompts?

<Accordion title="Prompt handling and retained metadata" defaultOpen>
  Token Factory processes prompts and generations in volatile memory; prompt
  caching can keep prompt data and KV caches in memory for several minutes.
  Prompt and generation data are not logged or stored persistently by default.
  Operational metadata, such as token counts, is retained. Read the
  [Data Security page](/security/data-security) for the full retention policy.
</Accordion>

## Does ZDR apply per model?

<Accordion title="Retention policy across models" defaultOpen>
  Token Factory's default zero data retention policy applies to its served
  open-weight models without a separate per-model opt-in. Operational metadata is
  retained. Review the [Data Security page](/security/data-security) before relying
  on ZDR for a workload.
</Accordion>

## What models do you support, and can I request one?

<Accordion title="Current models and model requests" defaultOpen>
  Token Factory currently serves `zai-org/GLM-5.3` and
  `deepseek-ai/DeepSeek-V4-Flash-0731`. Check the [model catalog](/models/overview)
  or `GET https://api.tokenfactory.corvex.cloud/v1/models` for current availability
  and capabilities. Send requests for other models through
  [Token Factory support](#how-do-i-contact-support).
</Accordion>

## What are the rate limits?

<Accordion title="Request limits and retries" defaultOpen>
  Token Factory allows **60 requests per minute per account**, shared across
  all your API keys.

  If you reach a limit, the API returns HTTP 429. Honor `Retry-After` when present;
  otherwise, retry with exponential backoff. Requests may also be temporarily limited when
  model capacity is busy.

  View your limits in the [dashboard](https://tokenfactory.corvex.cloud/app/settings/limits).
  For higher limits, contact [Token Factory support](#how-do-i-contact-support).
</Accordion>

## Do you offer embeddings?

<Accordion title="Embedding model availability" defaultOpen>
  Token Factory does not currently serve an embedding model. The presence of an
  embeddings endpoint does not mean a suitable model is available. Check the
  [model catalog](/models/overview) before choosing an endpoint and model.
</Accordion>

## How do I contact support?

<Accordion title="Open the Token Factory support form" defaultOpen>
  Sign in to the [Token Factory dashboard](https://tokenfactory.corvex.cloud/app)
  and select **Support** in the sidebar to open the **Contact us** form. The form
  uses your signed-in account to identify your request.

  If you cannot sign in or have a pre-sales question, email
  [tokenfactory-support@corvex.ai](mailto:tokenfactory-support@corvex.ai).
  For API errors, include the request ID, model, and error code. Do not send API
  keys or sensitive prompt content.
</Accordion>