#Hey,I have integrated ElevenLabs with my telephony system, where an agent should respond to incomi

1 messages · Page 1 of 1 (latest)

velvet krakenBOT
#
AI-generated Answer for payzinn

A 403 Forbidden error typically indicates an issue with permissions or access. Based on the information available:

  1. Plan Limitations: If you're on a Creator or other self-serve plan, real-time WebSocket streaming for custom integrations (like SIP) is not supported. This feature is only available with an Enterprise plan. You can explore upgrading to an Enterprise plan by submitting your requirements here.

  2. Authentication Requirements: Ensure that your API key has the necessary permissions and is being used correctly. If you're using a token generated via the /link API endpoint, note that tokens with the purpose "shareable_link" are not valid for WebSocket streaming.

  3. IP Restrictions: If you're encountering this error from a specific IP address, it could be due to IP restrictions. Check if your IP is blocked or restricted. If you're using a service like Ngrok, ensure it is not causing issues, as seen in a previous case where removing Ngrok from the allowlist resolved a similar problem.

  4. SIP-Specific Limitations: While there are no explicit mentions of SIP-specific restrictions, ensure that the audio format being used is compatible with ElevenLabs' requirements. PCM is often recommended for both input and output audio.

If these steps do not resolve the issue, you may need to upgrade to an Enterprise plan or contact the support team for further assistance.

deep lotus
#

Shoild I buy enterprise plan to have access on real-time websocket streaming?

velvet krakenBOT
#
AI-generated Answer for payzinn

Yes, real-time WebSocket streaming functionality for custom integrations, such as your SIP setup, is only available with the Enterprise plan. This explains the 403 Forbidden error you are encountering. The Creator plan and other self-serve plans (Free through Business) do not support this feature.

To gain access to real-time WebSocket streaming, custom server integrations, and advanced API features, you would need to upgrade to an Enterprise plan. You can learn more and submit your requirements on the Enterprise page.

blazing lagoon
#

can i change the user input audio format and tts output format depnding if its a webwidget? using dynamic variable?

velvet krakenBOT
#
AI-generated Answer for shawn_kerch

Currently, there isn't a way to dynamically change audio input/output format using variables at runtime. The only method available is to update the agent's audio input/output type by patching the agent settings before starting each conversation. This approach requires manual adjustments for each use case, such as switching between formats for a web widget or a calling agent.

For more details on how to patch agent settings, you can refer to the API documentation.

blazing lagoon
velvet krakenBOT
#
AI-generated Answer for shawn_kerch

Currently, ElevenLabs' conversational AI does not support dynamic variables for system prompts or runtime changes to audio input/output formats. To handle different audio formats (e.g., PCM for web browsers or u-law for calling agents), you need to manually update the agent's audio input/output settings by patching the agent configuration before starting each conversation.

For more details on how to update these settings, you can refer to the API documentation. Unfortunately, there isn't a built-in feature to dynamically switch formats based on context like a web widget at this time.

blazing lagoon
#

can i change the user input audio format and tts output format depnding if its a webwidget? using dynamic variable or using override?