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

# Ringg.AI

> Add and configure Ringg.AI Parrot Speech-to-Text (STT) in TrueFoundry's AI Gateway using Custom Endpoints

### Adding models

Add [Ringg.AI](https://www.ringg.ai/) Parrot STT to the AI Gateway using **Custom Endpoints**. Parrot is Ringg's speech-to-text model built for real-time voice agents, with strong support for Hindi, English, and code-mixed speech. Learn more on the [Parrot STT model page](https://www.ringg.ai/models/speech-to-text/v1) and in the [Parrot STT announcement](https://www.ringg.ai/blog/parrot-speech-to-text-api).

<Steps>
  <Step title="Create a Custom Endpoint provider account">
    In the TrueFoundry dashboard, go to **AI Gateway** → **Models** → **Custom Endpoints** and click **Add Custom Endpoint**.

    In **Configure Account**, set:

    * **Name**: any unique name (this becomes `{providerAccountName}` in the URL)
    * **Endpoint Type**: `None` (or any value you use for tracking)
    * **Header Auth**: keep **disabled**
  </Step>

  <Step title="Add a Ringg.AI STT endpoint">
    Click **Add Endpoint** (or **Add Endpoint** from the list view) and configure the integration:

    * **Display Name**: any name (this becomes `{endpointName}` in the URL)
    * **Base URL**: `https://prod-api.ringg.ai/stt/v1/transcriptions`

    Enable **Custom Headers** and add:

    * `x-api-key`: `<ringg-api-key>`

    Keep **Header Auth** disabled.

    <Frame caption="Endpoint settings — Base URL and upstream API key header">
      <img src="https://mintcdn.com/truefoundry/TNopa3M6_LjoCYde/images/ringgAI.png?fit=max&auto=format&n=TNopa3M6_LjoCYde&q=85&s=893cd388a57d5651df28da360222f2a2" alt="Custom Endpoint form for Ringg.AI STT with Base URL and x-api-key custom header" width="965" height="712" data-path="images/ringgAI.png" />
    </Frame>

    <Note>
      This **x-api-key** header is sent **from the gateway to Ringg.AI**. Your client should only send the **TrueFoundry API key** to the gateway.
    </Note>
  </Step>

  <Step title="Set access control and save">
    On the **Access Control** step, add the users/teams who can view/use this provider account, then **Save**.
  </Step>
</Steps>

### Inference

Once saved, call your endpoint through the gateway’s **proxy-api** path. URL shape and path rules are documented under [Custom Endpoints](/docs/ai-gateway/custom-endpoints#endpoint-structure).

### Supported APIs

| API                               | Endpoint                                          | Tracing | Cost Tracking                                               |
| --------------------------------- | ------------------------------------------------- | ------- | ----------------------------------------------------------- |
| [Speech-to-Text](#speech-to-text) | `/proxy-api/{providerAccountName}/{endpointName}` | **✅**   | <Icon icon="circle-xmark" iconType="regular" color="red" /> |

<AccordionGroup>
  <Accordion title="Speech-to-Text">
    **Before you start:** Replace `{GATEWAY_BASE_URL}` with your gateway base URL ([how to find it](/docs/ai-gateway/quick-start#gateway-base-url)) and set `TFY_API_KEY` to your TrueFoundry API key.

    Send a multipart form request with the audio file and optional transcription parameters (for example `language` and `enable_cap_punc`).

    ```python lines theme={"dark"}
    import json
    import os

    import requests

    GATEWAY_BASE_URL = "{GATEWAY_BASE_URL}"
    TFY_API_KEY = os.environ.get("TFY_API_KEY", "your-tfy-api-key")

    # providerAccountName = Custom Endpoint account name
    # endpointName = integration display name
    URL = f"{GATEWAY_BASE_URL}/proxy-api/{providerAccountName}/{endpointName}"

    with open("audio.mp3", "rb") as f:
        response = requests.post(
            URL,
            headers={"Authorization": f"Bearer {TFY_API_KEY}"},
            files={"file": f},
            data={
                "language": "hi",
                "enable_cap_punc": "false",
            },
            timeout=120,
        )

    response.raise_for_status()
    result = response.json()

    with open("response.json", "w", encoding="utf-8") as out:
        json.dump(result, out, ensure_ascii=False, indent=2)

    print("Saved response to response.json")
    ```

    <Info>
      **Support scope:** We do not support all APIs through this integration. Only **HTTPS** and **non-streaming** requests are supported. We will add more support over time, including additional endpoints on customer request. Send a request to [support@truefoundry.com](mailto:support@truefoundry.com).
    </Info>
  </Accordion>
</AccordionGroup>