#is there a way to connect with my custom AI endpoint ?

1 messages · Page 1 of 1 (latest)

near heraldBOT
#
AI-generated Answer for phungan

Yes, you can connect your custom AI endpoint to ElevenLabs' Conversational AI. To do this, you need to set up an API endpoint on your backend server that serves streamed responses following the OpenAI API specification. Once your server is ready, you can provide the server URL and necessary parameters in the ElevenLabs platform under the custom AI settings.

If you're experiencing issues with the connection, ensure your server is online and resolving correctly. Additionally, enabling the "Custom LLM extra body" option in the settings may help resolve certain problems. If you need further assistance, feel free to reach out to the support team.

clever vine
#

is it possible to call the agent into google meeting ?

near heraldBOT
#
AI-generated Answer for phungan

The provided information does not mention direct integration with Google Meet or connecting custom AI endpoints to it. However, you can generally integrate ElevenLabs' AI with other systems using APIs. For specific integration capabilities, you might need to explore the documentation of both platforms or consult with a developer to set up custom solutions.

If you have more details about your setup or requirements, feel free to share, and I can guide you further.

clever vine
#

i want to connect my aws bedrock agent to you through custom LLM and call the agent into google meeting

near heraldBOT
#
AI-generated Answer for phungan

To connect your AWS Bedrock agent to ElevenLabs through a custom LLM, you can integrate your custom AI endpoint as long as it adheres to the OpenAI chat completion format. This ensures compatibility with ElevenLabs' SDK. You can use any local or cloud-based LLM, including AWS Bedrock, provided it meets this requirement.

However, there is no information available about directly calling the agent into Google Meet. You may need to explore third-party integrations or custom solutions to bridge the agent's functionality with Google Meet. If you have more specific requirements, feel free to share them!

clever vine
#

i have my custom AI endpoint already with openai chat completion format, how should i connect to elevenlab, show me step by step

near heraldBOT
#
AI-generated Answer for phungan

To connect your custom AI endpoint to ElevenLabs using the Custom LLM feature, follow these steps:

1. Ensure Compatibility

Your custom AI endpoint must return responses in the OpenAI chat completion format. This ensures compatibility with ElevenLabs' SDK. If your endpoint is already set up in this format, you can proceed.

2. Set Up Your Server

Make sure your server is online and accessible via a public URL. If you're using tools like ngrok to expose your local server, ensure the URL is updated in ElevenLabs whenever it changes.

3. Configure Custom LLM in ElevenLabs

  • Log in to your ElevenLabs account.
  • Navigate to the Custom LLM settings.
  • Provide the server URL of your custom AI endpoint.
  • Add any necessary parameters, such as API keys or secrets, in the settings. If the API key field is grayed out, you need to add a secret in the settings to enable it.

4. Test the Integration

  • Initiate a test call from ElevenLabs to your custom endpoint.
  • Ensure the request and payload structure mirror the OpenAI chat completion structure. If you encounter errors like Cannot POST /elevenlabs/chat/completions, verify the endpoint URL and payload format.

5. Troubleshooting

  • If no requests are reaching your server, double-check the server URL and ensure it is active.
  • Verify that the authorization header is included in your requests.
  • If you are using a free ngrok account, remember that the URL changes each time you restart it. Update the URL in ElevenLabs accordingly.

6. Enable Additional Features

If needed, enable the "Custom LLM extra body" option in ElevenLabs settings. This can help resolve certain issues with the integration.

For further details, you can refer to ElevenLabs' documentation on integrating custom LLMs. Let me know if you need clarification on any step!

clever vine
#

can you chane the way read the number like 1000, instead of reading one zero zero zero zero, its reading one thousand

near heraldBOT
#
AI-generated Answer for phungan

To adjust how numbers like "1000" are read in ElevenLabs, you can modify the input text to influence the pronunciation. Here are some steps you can follow:

  1. Write Numbers as Words: If you want the AI to read "1000" as "one zero zero zero," you should write it out explicitly in your text prompt as "one zero zero zero." This ensures the AI interprets the input exactly as intended.

  2. Avoid Numeric Symbols: The AI may interpret numeric symbols like "1000" based on standard conventions (e.g., "one thousand"). Writing the numbers phonetically or as individual digits will help achieve the desired output.

  3. Use Text Normalization Settings (API): If you're using the ElevenLabs Text-to-Speech API, you can enable text normalization by adding "apply_text_normalization": "on" to your parameters. However, note that this setting may still default to standard number pronunciations unless explicitly written out.

If you continue to face issues, feel free to provide more details about your use case, and we can guide you further.

clever vine
#

should i write 350,000 for three hundred and fifty thousand or 350000

near heraldBOT
#
AI-generated Answer for phungan

To ensure accurate pronunciation by ElevenLabs' AI voices, it is recommended to write numbers in words rather than numerals. For example, instead of "350,000" or "350000," you should write "three hundred and fifty thousand." This approach helps the AI interpret the text correctly and produce the desired audio output.

clever vine
#

if i write $325, does the voice understand and read three hundred and twenty five dollars ? or it will read three two five dollars

near heraldBOT
#
AI-generated Answer for phungan

The AI may sometimes misinterpret numbers, especially when combined with currency symbols. To ensure accurate pronunciation, it's recommended to write out the number in full, such as "three hundred and twenty-five dollars" instead of "$325." This approach helps the AI produce the correct audio output.

clever vine
#

are you using rest api, webhook or websocket ?

near heraldBOT
#
AI-generated Answer for phungan

ElevenLabs Conversational AI primarily uses WebSocket connections for real-time interactions. For example, to start a conversation with an agent, you can connect to the WebSocket endpoint:

wss://api.elevenlabs.io/v1/convai/conversation?agent_id=<your-agent-id>

This connection should remain open throughout the duration of the call. If you're looking to integrate with other systems, such as Twilio or embedding a widget, WebSocket is the recommended approach.

Let me know if you need further clarification!

clever vine
#

im using custom llm, you require and endpoint api, is it rest api, webhook or websocket

near heraldBOT
#
AI-generated Answer for phungan

To integrate a custom LLM with ElevenLabs, the endpoint must be compatible with the OpenAI SDK and support the chat completion format. This typically involves using a REST API. Ensure your endpoint can handle POST requests structured like OpenAI's chat completions, including features such as parallel_tool_calls if required.

clever vine
#

can the endpoint api handle with websocket ?

near heraldBOT
#
AI-generated Answer for phungan

Yes, ElevenLabs provides a WebSocket API for real-time interactions. However, WebSocket support is currently limited to Text-to-Speech functionality and does not support real-time Speech-to-Speech use cases. If you're looking to integrate a custom LLM endpoint with WebSocket capabilities, you would need to ensure compatibility with ElevenLabs' WebSocket API for Text-to-Speech.

For more details on implementing WebSocket integrations, you can refer to the API documentation or consult with developers familiar with both platforms. If you require advanced features like real-time conversational AI WebSocket streaming, this functionality is only available under the Enterprise plan. You can explore Enterprise options at https://elevenlabs.io/enterprise.

clever vine
#

do i need to download anything for elevenlab ?

near heraldBOT
#
AI-generated Answer for phungan

You don’t need to download anything to useLabs. All features, including voice generation and editing, are accessible through a web browser. You can sign up for free and start using the platform directly online. If you're on a mobile device, you can also use a browser like Google Chrome to access the services.

clever vine
#

mpv not found necessary to stream audio, what is this error ?

near heraldBOT
#
AI-generated Answer for phungan

The error "mpv not found necessary to stream audio" typically means that the MPV media player, which is often used for audio or video streaming, is not installed or properly configured on your system. MPV is a lightweight, open-source media player that some applications rely on for streaming or playback.

To resolve this issue, you can try the following steps:

  1. Install MPV: Download and install MPV on your system. You can find installation instructions for your operating system on the official MPV website.

  2. Verify Installation: After installation, ensure that MPV is added to your system's PATH so it can be accessed from the command line or by other applications.

  3. Restart Your Application: If the error persists, restart the application or service that is trying to use MPV.

If you continue to encounter issues, you may need to check the specific application or service documentation for additional troubleshooting steps.

clever vine
#

where is. your server based

near heraldBOT
#
AI-generated Answer for phungan

ElevenLabs does not publicly disclose specific details about its server locations. If you are experiencing issues related to server access or geolocation, feel free to provide more details about your situation, such as your server's IP address and hosting provider, so the team can investigate further.

clever vine
#

i just want to know why your latency so low compare to other like VAPI ?

near heraldBOT
#
AI-generated Answer for phungan

The latency of ElevenLabs' Text-to-Speech (TTS) service can be influenced by several factors that help optimize performance:

  1. Turbo v2.5 Model: This model is specifically designed for low-latency tasks, ensuring faster audio generation.

  2. Streaming API: This API streams audio back during generation, reducing response times compared to regular endpoints.

  3. Input Streaming Websockets: By streaming text prompts in smaller chunks, audio is generated faster, minimizing latency.

  4. Enterprise Plan: Enterprise customers receive priority in the rendering queue, which significantly reduces latency.

  5. Voice Selection: Using default or synthetic voices typically results in lower latency compared to Professional Voice Clones.

  6. Connection Optimization: Techniques like reusing HTTPS sessions and limiting WebSocket connection closures reduce overhead and improve speed.

  7. Server Location: Leveraging servers closer to the US can improve latency for users outside the region, as network routing plays a role.

These optimizations collectively contribute to lower latency compared to other services. If you're experiencing latency issues or need further assistance, feel free to provide more details about your setup or voice model.

clever vine
#

i just want to know why your latency so low compare to other like VAPI ?