https://retellai.com/?utm_source=discord&utm_medium=seo logo
why there is delay for next node to start when I u...
# ❓・ai-evy-support
q
@AI Evy why there is delay for next node to start when I use custom function with "Wait for Result" off.
a
Thanks for tagging me @Q4. I'll look through relevant sources and get back to you shortly.
### Answer When Wait for Result is off, the transition timing depends on the Speak During Execution setting [(1) ]() : - If Speak During Execution is **on**: the agent will transition once done talking - If Speak During Execution is **off**: the agent should transition immediately after the function gets invoked (right upon entering the node) However, based on community reports, there may still be a delay even with both settings off—the function node appears to wait for some response from the custom function before transitioning, rather than running truly in the background [(2) ]() . This behavior has been observed specifically in chat mode vs. voice mode, where voice mode worked as expected . Possible workarounds: 1. Test on voice instead of chat if you're experiencing this in chat mode 2. Ensure both Wait for Result and Speak During Execution are disabled in the function node settings If you're still experiencing delays with these settings configured correctly, this may be unexpected behavior worth reporting.
​ To learn more, see the sources I considered: [[1] Function Node Overview](https://docs.retellai.com/build/conversation-flow/function-node) [[2] Asynchronous Function implementation node](https://discord.com/channels/1186424376300679209/1436155558192480317) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
q
@AI Evy Speak During Execution is off too and I am testing on real calls where delay occures
a
Thanks for the follow-up, @Q4. I'll get back to you soon.
### Answer --- Based on the documentation, when Wait for Result is off and Speak During Execution is off, the agent should transition immediately after the function gets invoked—right upon entering the node. [(1) ]() Additionally, if the user interrupts the agent, the transition can also happen once the user is done speaking. Since you're experiencing delays on real calls with both settings disabled, this behavior doesn't align with the documented expected behavior. The sources I have don't explain why this delay would occur in your scenario. Check out these potentially useful sources: - [Function Node Overview]() I'd recommend reaching out for further assistance: - Visit [Contact Sales]() - Visit [) Would you like me to provide a general answer based on my knowledge instead?
​ To learn more, see the sources I considered: [[1] Function Node Overview](https://docs.retellai.com/build/conversation-flow/function-node) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy what is the best practice for agent interaption definitions?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer ### Answer Here are best practices for configuring agent interruption handling: 1. Tune Interruption Sensitivity The
interruption_sensitivity
parameter controls how easily users can interrupt the agent. [(1) ]() - Value ranges from
[0,1]
- Lower values = harder for users to interrupt (agent waits longer/more words before yielding) - Higher values = easier for users to interrupt - Setting to
0
means the agent would never be interrupted For noisy environments, set it lower (e.g.,
0.8
) to reduce false interruptions from background speech and noise. [(2) ]() 2. Use Denoising Mode for Noisy Environments - Remove noise (default) : Removes background noise without impacting transcription accuracy - **Remove noise + background speech**: More aggressive mode for loud background speech (TV, construction sites) , but may affect transcription accuracy. Incurs $0.005/min surcharge. 3. Handle Interruptions in Your Prompt Community members have found success with prompt-based handling: [(3) ]() - When interrupted, have the agent pause and ask "Please go ahead" or "Did you say something?" - If the user says they didn't say anything, resume from where the agent stopped - If the user did speak, acknowledge it, repeat back what was said, then ask permission to continue 4. Use
no_interruption_allowed
for Critical Content
When using Custom LLM via WebSocket, you can set
no_interruption_allowed
to
true
in your response events so the agent won't be interrupted during important information. [(6) ]()
​ To learn more, see the sources I considered: [[1] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[2] Handle background speech & noise](https://docs.retellai.com/build/handle-background-noise) [[3] Discord message from @bajacheebo on 2025-01-07](https://discord.com/channels/1186424376300679209/1294891368505937952/1325992277759365161) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer ### Answer For a clinic receptionist AI voice agent, here are recommended configurations based on the documentation: Core Capabilities to Enable - Calendar syncing for appointment scheduling and management [(1) ]() - CRM synchronization with platforms like Salesforce or HubSpot to retrieve patient data - Knowledge base with dynamic knowledge retrieval and structured knowledge base injection for FAQs about services, hours, and policies Key Features - 24/7 availability to handle calls outside business hours [(2) ]() - Call routing based on intent or department [(1) ]() - Appointment booking capabilities - Message taking when staff are unavailable - Seamless escalation to human agents for complex inquiries , [(2) ]() Integrations to Consider The platform supports integrations with healthcare-specific systems including: - Jane App - Dentrix - OpenDental - Epic - ChiroTouch - eClinical Handling Unpredictable Callers Based on community experience building clinic voice agents, callers often mispronounce medical terms or use different names for the same procedure. [(3) ]() Recommendations include: - Analyzing existing call transcripts to identify common patterns - Starting with the most frequent call type (e.g., pricing inquiries) - Transferring edge cases to human agents initially, then expanding automation over time Security For healthcare, Retell AI offers HIPAA compliance, end-to-end encryption, and granular data retention rules. [(1) ](https:/
/www.retellai.com/blog/transform-your-front-desk-with-retell-ai-receptionist)
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
​ To learn more, see the sources I considered: [[1] Transform Your Front Desk With Retell AI Receptionist | Retell AI | Retell AI](https://www.retellai.com/blog/transform-your-front-desk-with-retell-ai-receptionist) [[2] Create Realistic Al Voice Agents for Receptionists](https://www.retellai.com/use-cases/receptionists) [[3] Discord message from @pa2791 on 2025-04-15](https://discord.com/channels/1186424376300679209/1294891368505937952/1361798373505499146) ​
s
@User and voice mode, temperature, speed and volume?
@AI Evy and voice mode, temperature, speed and volume?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer ### Answer For a clinic receptionist, here are recommended voice settings: Voice Model Options [(1) ]() Available voice models include: -
eleven_turbo_v2
,
eleven_flash_v2
,
eleven_turbo_v2_5
,
eleven_flash_v2_5
,
eleven_multilingual_v2
(ElevenLabs) -
sonic-2
,
sonic-3
,
sonic-3-latest
,
sonic-turbo
(Cartesia) -
tts-1
,
gpt-4o-mini-tts
(OpenAI) -
speech-02-turbo
(MiniMax) Voice Temperature - Range:
[0, 2]
- Lower values = more stable/consistent voice - Higher values = more variant speech generation - Default:
1
- Recommendation for clinic: Keep at
1
or slightly lower (e.g.,
0.8
) for consistent, professional-sounding responses Voice Speed - Range:
[0.5, 2]
- Lower values = slower speech - Higher values = faster speech - Default:
1
- Recommendation for clinic:
1
or slightly slower (e.g.,
0.9
) to ensure clarity, especially for appointment details Volume - Range:
[0, 2]
- Lower values = quieter speech - Higher values = louder speech - Default:
1
- Recommendation for clinic:
1
is typically appropriate; adjust based on caller feedback Note: Voice temperature currently only applies to ElevenLabs (
11labs
) voices.
​ To learn more, see the sources I considered: [[1] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy what is the best practice for user interaption definitions?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer Here are best practices for configuring user interruption handling: 1. Configure Interruption Sensitivity The
interruption_sensitivity
parameter controls how sensitive the agent is to user interruptions. [(1) ]() - Range:
[0, 1]
- Lower values = takes longer/more words for user to interrupt the agent - Higher values = easier for user to interrupt the agent - When set to
0
, agent would never be interrupted - Default:
1
For noisy environments or to reduce false interruptions from background speech, set it lower (e.g.,
0.8
) . [(2) ]() 2. Set Denoising Mode Two options available: - Remove noise (default) : Removes background noise with nearly no waveform distortion, no impact on speech-to-text accuracy - **Remove noise + background speech**: More aggressive mode for loud background speech (TV, construction sites) . May distort waveform and lower transcription accuracy in some cases. Incurs $0.005/min surcharge. 3. Handle Interruptions via Prompt Community members have found success with prompt-based handling: [(3) ]() , [(4) ]() - When interrupted, have the agent pause and say "Please go ahead" or "Did you say something?" - If user responds "No, I didn't say anything": Resume from where the agent stopped - If user did speak: Acknowledge with "Thank you for sharing that," repeat back what was said, then ask "May I share what I was saying before we continue?" - Always ensure the agent resumes the original point instead of jumping ahead [(5) ]() **4. Additio
nal Tips** - Lower interruption sensitivity also helps when talking to elderly users who may speak slower [(6) ]() - Have the agent ask if it's a good time to talk first, especially in noisy environments [(3) ]()
​ To learn more, see the sources I considered: [[1] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[2] Handle background speech & noise](https://docs.retellai.com/build/handle-background-noise) [[3] Discord message from @mattybgoood on 2025-01-07](https://discord.com/channels/1186424376300679209/1294891368505937952/1326210671008022663) [[4] Discord message from @bajacheebo on 2025-01-07](https://discord.com/channels/1186424376300679209/1294891368505937952/1325992277759365161) [[5] Discord message from @qcumbers1_cc on 2025-01-09](https://discord.com/channels/1186424376300679209/1294891368505937952/1326872454764695602) [[6] Step 2: Configure the basic settings](https://docs.retellai.com/build/single-multi-prompt/configure-basic-settings) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy what prmopts are also best practice for making sure the agent listens to the end and does not interrupt the user.
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer Here are prompt-based best practices to ensure the agent listens fully and doesn't interrupt users: Interruption Handling Prompt Example [(1) ]() Add this to your prompt:
Copy code
## Interruption Handling:
- If a customer makes a noise or interrupts while you are speaking:  
  - Pause and say:  
    - "Please go ahead." or "Did you say something?"  
  - **If they respond with**:  
    - **"No, I didn't say anything":** Resume from where you stopped in your sentence.  
    - **"Yes, I was trying to say...":**  
      - Listen carefully.  
      - Acknowledge with: "Thank you for sharing that."  Then repeat back what they said to confirm with them that you heard them.
      - Politely ask: "May I share what I was saying before we continue?"  

## Maintaining Context:
- If interrupted, always ensure you resume the original point instead of jumping ahead.  
- Avoid skipping key parts of the explanation unless the customer explicitly requests it.
Additional Prompt Tips - Include backchanneling phrases like "That's great!", "Wonderful!", "Good to know!" to acknowledge user input - Have the agent ask if it's a good time to talk first, especially in noisy environments Combine with Settings Lower the
interruption_sensitivity
parameter (e.g.,
0.8
or lower) to make it harder for background noise to trigger interruptions. [(2) ]() , [(3) ]() For extremely noisy environments, you may need to experiment with even lower settings, though this will also make it harder for users to intentionally interrupt. [(2) ]()
​ To learn more, see the sources I considered: [[1] Discord message from @bajacheebo on 2025-01-07](https://discord.com/channels/1186424376300679209/1294891368505937952/1325992277759365161) [[2] Handle background speech & noise](https://docs.retellai.com/build/handle-background-noise) [[3] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy what is the most recommended Responsiveness?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer The
responsiveness
parameter controls how responsive the agent is during conversations. [(1) ]() , [(2) ]() , [(3) ]() Configuration Details: - Range:
[0, 1]
[(1) ]() - Lower values = less responsive agent (waits more, responds slower) - Higher values = faster exchanges (responds when it can) - Default:
1
Recommendation for a Clinic Receptionist: The documentation doesn't specify a "most recommended" value—it depends on your use case. However, based on the parameter behavior: - **Default `1`**: Best for natural, quick conversational exchanges where you want the agent to respond promptly - **Lower values (e.g., `0.7`–`0.9`) **: Better if you want the agent to wait longer before responding, giving users more time to finish speaking For a clinic receptionist handling appointment details or elderly callers who may speak slower, you might consider lowering responsiveness slightly (e.g.,
0.8
or
0.9
) to ensure the agent doesn't respond too quickly before the caller finishes. Combine this with lower
interruption_sensitivity
for best results. [(4) ]()
​ To learn more, see the sources I considered: [[1] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) [[2] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[3] Get Voice Agent](https://docs.retellai.com/api-references/get-agent) [[4] Handle background speech & noise](https://docs.retellai.com/build/handle-background-noise) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , give me the best practice prompt for not getting into customer's speaking
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer Here's a best practice prompt for ensuring the agent doesn't interrupt while the customer is speaking: [(1) ]()
Copy code
## Interruption Handling:
- If a customer makes a noise or interrupts while you are speaking:  
  - Pause and say:  
    - "Please go ahead." or "Did you say something?"  
  - **If they respond with**:  
    - **"No, I didn't say anything":** Resume from where you stopped in your sentence.  
    - **"Yes, I was trying to say...":**  
      - Listen carefully.  
      - Acknowledge with: "Thank you for sharing that."  Then repeat back what they said to confirm with them that you heard them.
      - Politely ask: "May I share what I was saying before we continue?"  

## Maintaining Context:
- If interrupted, always ensure you resume the original point instead of jumping ahead.  
- Avoid skipping key parts of the explanation unless the customer explicitly requests it.
Additional prompt guidance for long user responses: [(2) ]() When expecting extended user input (like describing an incident) , include explicit instructions:
Copy code
## Golden Rule
As you listen to the user's description, you will focus on understanding the context and content of their statement. Only at the end of their statement should you move on to the next step. Under no circumstance should you move to the next step while the user is still describing. If you are not sure whether or not the user has completed their description, you will ask in a polite manner to confirm if you can proceed, before moving to the next step. Never assume that the user has completed their account based on a pause alone; pauses in speech don't necessarily mean the user has completed.
Combine with settings: - Lower
interruption_sensitivity
(e.g.,
0.5
or lower) [(3) ]() , [(4) ]() - Lower
responsiveness
(e.g.,
0.8
) to make the agent wait longer before responding [(3) ]() , [(4) ]()
​ To learn more, see the sources I considered: [[1] Discord message from @mattybgoood on 2025-01-07](https://discord.com/channels/1186424376300679209/1294891368505937952/1326210671008022663) [[2] Agent Interruption on Long Response](https://discord.com/channels/1186424376300679209/1295343988236353628) [[3] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[4] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy I need a prompt that makes the agent say the sentence to the end without being interrupted and so that the conversation is fluent and flowing.
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer To ensure the agent completes sentences without interruption while maintaining fluent conversation, combine these approaches: 1. Use Block Interruptions Setting In Conversation Flow, enable Block Interruptions on nodes where the agent must finish speaking. [(1) ]() 2. Configure Agent Settings Set
interruption_sensitivity
to
0
to prevent the agent from ever being interrupted. [(2) ]()
Copy code
json
"interruption_sensitivity": 0
3. Prompt for Graceful Recovery Add this to your prompt to handle any accidental interruptions while keeping flow natural: [(3) ]()
Copy code
## Backchanneling:
- When the customer responds to a question or shares information, use one of the following backchanneling phrases to acknowledge their input and keep the conversation engaging:
  1. "That's great!"
  2. "Wonderful!"
  3. "I love hearing that!"
  4. "Fantastic!"
  5. "Good to know!"
  6. "Interesting!"

- Alternate between these phrases throughout the conversation to maintain a natural tone.

## Interruption Handling:
- If a customer makes a noise or interrupts while you are speaking:  
  - Pause and say:  
    - "Please go ahead." or "Did you say something?"  
  - **If they respond with**:  
    - **"No, I didn't say anything":** Resume from where you stopped in your sentence.  
    - **"Yes, I was trying to say...":**  
      - Listen carefully.  
      - Acknowledge with: "Thank you for sharing that." Then repeat back what they said to confirm.
      - Politely ask: "May I share what I was saying before we continue?"  

## Maintaining Context:
- If interrupted, always ensure you resume the original point instead of jumping ahead.  
- Avoid skipping key parts of the explanation unless the customer explicitly requests it.
Note on Keypad Interruptions Be aware that Block Interruptions currently only applies to voice input—keypad presses can still interrupt. This has been reported as a feature request. [(4) ]()
​ To learn more, see the sources I considered: [[1] Conversation Node](https://docs.retellai.com/build/conversation-flow/conversation-node) [[2] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[3] Discord message from @bajacheebo on 2025-01-07](https://discord.com/channels/1186424376300679209/1294891368505937952/1325992277759365161) [[4] Prompt gets interrupted by early keypress before finishing playback](https://discord.com/channels/1186424376300679209/1420140310058631321) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy is there "Post-match endpointing" at the settings?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer I don't have documentation that specifically mentions a setting called "Post-match endpointing" in Retell AI's agent configuration. The available sources discuss related settings like
interruption_sensitivity
and
responsiveness
, but do not reference "Post-match endpointing" by that name. Check out these potentially useful sources: - [Contact Sales]() - [) Would you like me to provide a general answer based on my knowledge instead?
message has been deleted
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , please explain what is the best way at conversational flow to check id structure\logic ID Integrity Check (CRITICAL): When the user provides an ID number, you MUST verify its validity (Israeli ID checksum and 9 digits). - Israeli ID Validation Protocol: When a user provides an ID number, you MUST follow this internal logic before confirming: Normalization: If the input has 7 or 8 digits, add a leading '0'. If it has fewer than 8 or more than 9, ask the user to repeat it. Step-by-Step Calculation (Luhn): - Multiply digits by weights: [1, 2, 1, 2, 1, 2, 1, 2, 1].For any result > 9, sum the individual digits (e.g., $18 \rightarrow 1+8=9$).Sum all final values.Verification: The ID is valid ONLY if the sum is divisible by 10 (e.g., 60, 70, 80).Error Handling: If the ID is invalid, do not say "It's illegal." Instead, say: " The ID number is wrong please try again
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer For validating ID structure/logic like an Israeli ID checksum in Conversation Flow, the best approach is to use a *custom function (tool) * that performs the validation externally, then returns the result to the flow. Recommended Approach: 1. Create a Custom Function that receives the ID number, performs the Luhn checksum validation logic, and returns whether it's valid or invalid. [(1) ]() 2. Use a Component to encapsulate the ID verification sub-flow for reusability across agents. Components let you build once and reuse across flows while keeping your main canvas clean. 3. Structure the flow: - **Conversation Node**: Ask for the ID number - **Function Node**: Call your validation function with the ID - **Conditional transitions**: Based on the function result, either proceed (valid) or loop back with an error message (invalid) Why use a custom function: - The Luhn algorithm calculation (multiplying by weights, summing digits, checking divisibility by 10) is complex logic better handled by external code than prompt instructions - Prompts can hallucinate or miscalculate; a function guarantees accuracy [(2) ](

▾

) For the error handling prompt, use something like:
Copy code
If the ID validation returns invalid, say: "The ID number appears to be incorrect. Could you please repeat it?"
The tools need to be defined within the component and will not be visible outside at agent level. [(1) ]()
​ To learn more, see the sources I considered: [[1] Components](https://docs.retellai.com/build/conversation-flow/components) [[2] Master The Future of Voice AI - Conversation Design & Agent Persona](

https://www.youtube.com/watch?v=42V0bDQIsb0▾

) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , i would like to avoid using http request and do the calc inside the flow, what will be the best way?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer Conversation Flow does not support native in-flow calculations or scripting logic—all function execution requires external HTTP requests via custom functions. [(1) ]() Your options within Conversation Flow: 1. **Prompt-based validation (limited reliability) **: You could include the Luhn algorithm logic in your prompt instructions and ask the LLM to perform the calculation. However, this approach is not recommended for critical validation because LLMs can hallucinate or miscalculate mathematical operations. [(2) ]() 2. **Equation-Based Transitions**: Conversation Flow supports equation-based transitions for conditional logic (e.g., checking if a dynamic variable equals a value) , but this is designed for simple comparisons—not complex multi-step algorithms like Luhn checksum calculations. [(3) ]() Why external functions are recommended: Custom functions are the intended mechanism for executing logic like ID validation because they guarantee accuracy. The function receives parameters, your backend performs the calculation, and returns the result. [(1) ]() Bottom line: There is no built-in way to perform complex calculations like the Israeli ID Luhn checksum purely within Conversation Flow without an HTTP request. If you need guaranteed accuracy for ID validation, a custom function calling an external endpoint remains the best practice. Would you like guidance on setting up a lightweight validation endpoint?
​ To learn more, see the sources I considered: [[1] Custom Function](https://docs.retellai.com/build/conversation-flow/custom-function) [[2] Prompt Engineering Guide](https://docs.retellai.com/build/prompt-engineering-guide) [[3] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , can i use the logic node in this case? and if yes explain how
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer The Logic Split Node is designed for branching based on conditions, but it **cannot perform calculations**—it only evaluates existing variable values. [(1) ]() What Logic Split Node does: - Immediately evaluates conditions when the agent enters the node - Branches to different destination nodes based on whether conditions are met - The agent does not speak in this node, and time spent is minimal - Useful for branching based on dynamic variables Why it won't work for Israeli ID validation: The Logic Split Node checks conditions like "if
{{variable}}
equals 'true'" or uses equation-based transitions for simple comparisons. , [(2) ]() For your Luhn checksum validation, you would need to: 1. Multiply 9 digits by weights 2. Sum digits of results > 9 3. Sum all values 4. Check if divisible by 10 This multi-step mathematical calculation cannot be performed within a Logic Split Node—it can only evaluate pre-existing values, not compute new ones. Your realistic options: 1. Custom function (recommended) : External endpoint performs the calculation and returns valid/invalid 2. Prompt-based (unreliable) : Ask the LLM to calculate in a Conversation Node, but LLMs often miscalculate math If you had a function that returned
{{id_valid}}
as
true
or
false
, then you could use a Logic Split Node to branch based on that result. [(3) ]()
​ To learn more, see the sources I considered: [[1] Logic Split Node](https://docs.retellai.com/build/conversation-flow/logic-split-node) [[2] Logic split node stuck](https://discord.com/channels/1186424376300679209/1435621223294959616) [[3] Simple check / decision not working](https://discord.com/channels/1186424376300679209/1399851200702316584) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy ת How do I speed up conversational flow?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer --- Here are ways to speed up your Conversation Flow agent: 1. Adjust Model Temperature Lower the
model_temperature
parameter for faster, more deterministic responses. [(1) ]()
Copy code
json
"model_temperature": 0.7
2. Use Different Models Per Node Since pricing and performance are calculated based on time spent in each node, you can optimize by using cheaper/faster models for simple routing nodes and premium models only for complex interactions. [(2) ]() 3. Split Complex Nodes If a node contains too many instructions, the agent may struggle to follow them consistently. Split complex nodes into multiple simpler nodes for more predictable behavior. [(3) ]() 4. Simplify Transition Conditions Complex transition conditions can slow down decision-making. Break down complex conditions into multiple simpler ones. 5. Review Node Structure Every node defines a small set of logic, and transition conditions determine which node to transition to. Keep each node focused on a specific task to reduce processing time. [(2) ]() 6. Check Latency in Dashboard You can view the estimated latency and cost of the agent within the Dashboard. Modify settings there—all changes are auto-saved.
​ To learn more, see the sources I considered: [[1] Create Conversation Flow](https://docs.retellai.com/api-references/create-conversation-flow) [[2] Conversation Flow Overview](https://docs.retellai.com/build/conversation-flow/overview) [[3] Debug guide](https://docs.retellai.com/build/conversation-flow/debug-guide) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy Does conversational flow support Pause Before Speaking?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer Yes, Conversation Flow supports Pause Before Speaking. In the Dashboard: Under Call Settings in the global settings, you can configure "Pause before speaking" — for the beginning of the call, if the agent speaks first, it will wait for the configured duration before speaking. This is useful to handle scenarios when the user is still picking up the phone. [(1) ]() Via API: You can set
begin_message_delay_ms
when creating or updating an agent. [(2) ]() The
begin_after_user_silence_ms
parameter is also available in the Conversation Flow API — if set, the AI will begin the conversation after waiting for the user for the duration (in milliseconds) specified. This only applies if the agent is configured to wait for the user to speak first. [(3) ]() Tip: Enable Pause Before Speaking to avoid cutting users off at the start of calls. [(4) ]()
​ To learn more, see the sources I considered: [[1] Step 1: Configure global settings](https://docs.retellai.com/build/conversation-flow/global-setting) [[2] pause before speaking in API](https://discord.com/channels/1186424376300679209/1328429685474070589) [[3] Update Conversation Flow](https://docs.retellai.com/api-references/update-conversation-flow) [[4] AI Interruptions](https://discord.com/channels/1186424376300679209/1359363109822922922) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , how do i use end call node ("ending") so it will end and close the call
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer The End Node terminates the call the moment the agent enters it. [(1) ]() How to use it: 1. Add an End Node from the left sidebar to your canvas [(2) ]() 2. Connect it as the destination from your final conversation node 3. The call ends immediately when the agent transitions to this node [(1) ]() Node Settings: - **Speak During Execution**: Enable this to have the agent say a farewell message (like "Goodbye, have a nice day") while ending the call. You can choose between
Prompt
or
Static Sentence
. - **Global Node**: You can set the End Node as a global node so it's accessible from any point in the conversation Important tip for smooth endings: If you want a brief pause before the call ends (so it doesn't feel abrupt) , create a Conversation Node before your End Node with your farewell message and enable Skip Response, then connect it to the End Node. [(3) ]() You can also set the End Node as a global node combined with max call duration and max silence duration settings to ensure the agent eventually hangs up even if the end condition isn't triggered naturally. [(4) ]()
​ To learn more, see the sources I considered: [[1] End Node](https://docs.retellai.com/build/conversation-flow/end-node) [[2] Node Overview](https://docs.retellai.com/build/conversation-flow/node) [[3] Can I add a 1-second delay to "End Call" node?](https://discord.com/channels/1186424376300679209/1355665071363067916) [[4] Discord message from @mike_56558 on 2025-03-14](https://discord.com/channels/1186424376300679209/1294891368505937952/1350192812179849277) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , how do i use end call node ("ending") so it will end and close the call at conversational node
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer The End Node terminates the call the moment the agent enters it—it does not have any edges. [(1) ]() To connect from a Conversation Node to the End Node: 1. Add an End Node from the left sidebar to your canvas [(2) ]() 2. Click and hold the circle at the bottom of your Conversation Node to start a line that connects it to the End Node 3. Add transition conditions on the edge (e.g., when the user says goodbye or the conversation is complete) 4. The call ends immediately when the agent transitions to this node [(1) ]() For a smooth ending with a farewell message: Enable Speak During Execution in the End Node settings—this lets you write instructions for the agent to generate an utterance like "Goodbye, have a nice day" while the call is ending. You can choose between
Prompt
or
Static Sentence
. If "Speak During Execution" doesn't work reliably: A workaround is to move the farewell logic to your prompt instead of relying on the node setting. Instruct the agent to output the farewell sentence first, then immediately trigger the end call—without waiting for user input between the speech and the hangup. [(3) ]() Best practice for reliability: You can set the End Node as a Global Node so it's accessible from any point in the conversation. Combined with max call duration and max silence duration settings, this ensures your agent hangs up eventually. [(4) ]() , [(5) ]()
​ To learn more, see the sources I considered: [[1] End Node](https://docs.retellai.com/build/conversation-flow/end-node) [[2] Node Overview](https://docs.retellai.com/build/conversation-flow/node) [[3] Speak During Execution not working for End Call node](https://discord.com/channels/1186424376300679209/1451006517909258352) [[4] Discord message from @sanava.ai on 2025-03-14](https://discord.com/channels/1186424376300679209/1294891368505937952/1350212730593873941) [[5] Discord message from @mike_56558 on 2025-03-14](https://discord.com/channels/1186424376300679209/1294891368505937952/1350192812179849277) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , how do you suggest me to configure the transfer_call function so it will not transfer the call to human after work hours?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer To prevent call transfers outside work hours, use dynamic variables combined with conditional logic in your Conversation Flow or prompt. Recommended Approach: 1. Pass work hours status via dynamic variable When initiating the call, inject a dynamic variable like
{{is_within_work_hours}}
that your backend calculates based on current time. [(1) ]() 2. Use conditional logic before transfer In Conversation Flow, add a Logic Split Node before the Transfer Node to check the
{{is_within_work_hours}}
variable. Only route to the transfer node if the value is
true
. 3. Update your prompt with transfer conditions Include instructions in your prompt such as:
Copy code
If the user requests to speak with a human and {{is_within_work_hours}} is true, use the transfer_call tool to transfer the conversation to a human agent. If {{is_within_work_hours}} is false, inform the user that our team is currently unavailable and offer to take a message or schedule a callback.
4. Alternative: Handle time logic in a custom function Create a custom function that checks the current time against your business hours and returns whether transfer is allowed. Use the function result to conditionally trigger the transfer. Key point: The transfer_call function itself doesn't have built-in time-based restrictions—you need to implement the business hours logic externally (via dynamic variables or a custom function) and use that to control when the transfer is triggered.
​ To learn more, see the sources I considered: [[1] Transfer call](https://docs.retellai.com/build/single-multi-prompt/transfer-call) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer I don't have enough context to see what you're referring to—your message mentions "here" but I can't see an image or screenshot. Could you describe what screen or setting you're looking at? For example: - Is this a Transfer Node configuration in Conversation Flow? - A Logic Split Node for checking work hours? - A custom function setup? If you're configuring a transfer to not happen after work hours, you'll need to set up either: 1. A Logic Split Node that checks a dynamic variable like
{{is_within_work_hours}}
before routing to the Transfer Node [(1) ]() 2. Transition conditions on the edge leading to your transfer node that evaluate whether it's within business hours Please share more details or describe what you're seeing, and I can provide specific guidance.
​ To learn more, see the sources I considered: [[1] Retell AI’s Advanced Conversation Flow | Retell AI](https://www.retellai.com/blog/unlocking-complex-interactions-with-retell-ais-conversation-flow) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , where do i set adynamnic variable?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer You can set dynamic variables in multiple places: *1. Agent-Level Default Values (Dashboard) * Configure default values for dynamic variables at the agent level. Default variables serve as a fallback and will only be used when specific variables aren't included in the call request. [(1) ]() *2. For Outbound Calls (API) * When using the [Create Phone Call]() API, set your variables in the
retell_llm_dynamic_variables
field:
Copy code
json
{
  "user_name": "John Smith",
  "product_name": "Premium Plan",
  "account_status": "active"
}
*3. For Inbound Calls (Webhook) * Use the inbound call webhook response with the
call_inbound
field containing `dynamic_variables`: [(2) ]()
Copy code
json
{
  "call_inbound": {
    "dynamic_variables": {
      "caller_name": "value"
    }
  }
}
4. For HubSpot Integration Pass data using the Dynamic Variables field in JSON format when configuring the call form. Ensure all values are surrounded by quotes. [(3) ]() 5. Test Calls from Dashboard You can enter dynamic variables when sending a test call directly from the platform. [(4) ]() Important: For inbound webhooks, make sure to use the correct format with
call_inbound
wrapping—using
retell_llm_dynamic_variables
directly in the webhook response will cause an error. [(2) ]()
​ To learn more, see the sources I considered: [[1] Dynamic Variables](https://docs.retellai.com/build/dynamic-variables) [[2] Agent not picking up dynamic variables](https://discord.com/channels/1186424376300679209/1395530229640073329) [[3] HubSpot Integration](https://docs.retellai.com/integrations/hubspot) [[4] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy where do i configure the {{is_within_work_hours}}?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer
{{is_within_work_hours}}
is a **custom dynamic variable you create**—it's not a built-in system variable. You need to calculate it externally and pass it to Retell. Where to configure it: For Outbound Calls (API) : Set it in the
retell_llm_dynamic_variables
field when calling the [Create Phone Call]() API: [(1) ]()
Copy code
json
{
  "is_within_work_hours": "true"
}
For Inbound Calls (Webhook) : Return it in your inbound call webhook response: [(2) ]()
Copy code
json
{
  "call_inbound": {
    "dynamic_variables": {
      "is_within_work_hours": "true"
    }
  }
}
For Agent-Level Defaults: Configure default values in the agent settings in the Dashboard. These serve as a fallback when variables aren't included in the call request. [(1) ]() Important: All values must be strings (e.g.,
"true"
not
true
) . Your backend logic should calculate whether it's within work hours before initiating the call or responding to the inbound webhook, then pass the appropriate value.
​ To learn more, see the sources I considered: [[1] Dynamic Variables](https://docs.retellai.com/build/dynamic-variables) [[2] Dynamic Variables](https://discord.com/channels/1186424376300679209/1356680357084135558) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy where do i configure the {{is_within_work_hours}} at conversational flow?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer In Conversation Flow, you don't configure
{{is_within_work_hours}}
directly within the flow editor—dynamic variables like this are passed into the call externally (via API for outbound calls or webhook for inbound calls) as explained in my previous response. How to use it within Conversation Flow: Once the variable is passed into the call, you can use a Logic Split Node to branch based on its value. [(1) ]() 1. Add a Logic Split Node before your Transfer Node [(2) ]() 2. Configure a condition that checks if
{{is_within_work_hours}}
equals
"true"
[(1) ]() 3. Connect the matching path to your Transfer Node 4. Connect the else path (when outside work hours) to a Conversation Node that offers alternatives like taking a message Important note from community feedback: Logic Split Nodes evaluate existing variable values—they cannot calculate or create values. The variable must already be populated when the call reaches that node. There have been reports of Logic Split Nodes defaulting to the "else" path when variables aren't properly populated before reaching the node. [(3) ]() Make sure your
{{is_within_work_hours}}
variable is being passed correctly at call initiation (outbound API or inbound webhook) before the flow reaches the Logic Split Node.
​ To learn more, see the sources I considered: [[1] Logic Split Node](https://docs.retellai.com/build/conversation-flow/logic-split-node) [[2] Node Overview](https://docs.retellai.com/build/conversation-flow/node) [[3] Help with Logic Split Nodes (and an example of a custom variable being used would be lovely!)](https://discord.com/channels/1186424376300679209/1358825261986087126) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , where do i setup "Post-response-delay" at single prompt?\
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer I don't have specific information about a "Post-response-delay" setting for single/multi-prompt agents in the available sources. For single/multi-prompt agents, the Speech Settings section includes options like: [(1) ]() - **Responsiveness**: how responsive the agent is—reducing responsiveness by 0.1 adds 0.5 seconds of agent wait time - **Pause before speaking**: for the beginning of the call, if agent speaks first, it will wait for the configured duration before speaking If you're looking to add delay between responses, adjusting Responsiveness may help—lowering it makes the agent respond slower. Could you clarify what specific behavior you're trying to achieve? That would help me point you to the right setting. Check out these potentially useful sources: - [Step 2: Configure the basic settings]()
​ To learn more, see the sources I considered: [[1] Step 2: Configure the basic settings](https://docs.retellai.com/build/single-multi-prompt/configure-basic-settings) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , what will be the most effective agent ai model for hebrew?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer I don't have specific information about which LLM performs best for Hebrew in the available sources. The available models in Retell AI include: [(1) ]() -
gpt-4.1
,
gpt-4.1-mini
,
gpt-4.1-nano
-
gpt-5
,
gpt-5.1
,
gpt-5.2
,
gpt-5-mini
,
gpt-5-nano
-
claude-4.5-sonnet
,
claude-4.5-haiku
-
gemini-2.5-flash
,
gemini-2.5-flash-lite
,
gemini-3.0-flash
For speech-to-speech models:
gpt-4o-realtime
,
gpt-4o-mini-realtime
,
gpt-realtime
,
gpt-realtime-mini
Many LLMs are capable of handling multiple languages, enabling global voice AI applications. [(2) ]() Recommendation: Since the documentation doesn't specify Hebrew performance benchmarks, I'd suggest testing a few models (like
gpt-4.1
or
claude-4.5-sonnet
) with Hebrew conversations to evaluate accuracy and naturalness for your use case. You may also want to reach out to Retell support for specific multilingual guidance. Check out these potentially useful sources: - [Choosing the Best LLM for Your AI Voice Agents]() - [Contact Sales]() Would you like a general answer based on broader LLM knowledge about Hebrew language support?
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
​ To learn more, see the sources I considered: [[1] Update Retell LLM](https://docs.retellai.com/api-references/update-retell-llm) [[2] Large Language Model (LLM)](https://www.retellai.com/glossary/large-language-model-llm) ​
s
@AI Evy , Do you have any idea why tests I did in chat and then in a web call and were fine then didn't work in a phone call? Aren't they supposed to reflect exactly the same thing?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer Chat, web calls, and phone calls can behave differently due to several factors: Audio Processing Differences: - Phone Calls use different telephony providers with different audio codecs. Retell's telephony integrations handle encoding/decoding internally. [(1) ]() - Web Calls use the frontend web JS SDK which captures audio in PCM format and sends it to the backend for processing. Turn-Taking Behavior: Phone calls involve additional complexity with turn-taking. Retell AI's turn-taking model uses Voice Activity Detection (VAD) and predictive timing to determine when to listen or speak. [(2) ]() Traditional VAD systems can misinterpret pauses or background noise as the end of a speaker's turn, leading to interruptions or awkward silences. These issues may be more pronounced over phone lines due to audio quality variations, background noise, and latency differences compared to web calls. Potential Causes: 1. Audio quality/noise on phone lines affecting speech recognition accuracy 2. Latency differences between web and telephony 3. Turn-taking timing being affected by phone line characteristics 4. Telephony provider configuration differences Troubleshooting suggestions: - Check if the issue is with speech recognition (ASR) or the agent's responses - Review call transcripts to see if the agent heard the user correctly - Consider enabling Denoising Mode which can remove background noise and unwanted speech [(3) ]() Could you share more details about what specifically "didn't work"—was it the agent not understanding, responding incorrectly, or something else?
​ To learn more, see the sources I considered: [[1] Audio Basics](https://docs.retellai.com/knowledge/audio-basics) [[2] Retell AI's Turn-Taking Model for Seamless Calls | Retell AI](https://www.retellai.com/blog/how-retell-ais-turn-taking-model-ensures-seamless-calls) [[3] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , I have interruptions at the beginning of every sentence the agent says. What should I set? I have Pause Before Speaking: 0.4s
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer If the agent is getting interrupted at the beginning of every sentence, here are the key settings to adjust: 1. Lower Interruption Sensitivity [(1) ]() This controls how fast the agent gets interrupted by user interruptions. Set it lower if you want the agent to be more resilient to background speech or user interruptions. 2. Adjust Responsiveness This controls how responsive the agent is. Reducing responsiveness by 0.1 adds 0.5 seconds of agent wait time. However, don't set it too low or call quality may suffer. [(2) ]() 3. Pause Before Speaking [(1) ]() This setting is specifically for the beginning of the call when the agent speaks first—it makes the agent wait for the configured duration before speaking. Note: This setting may not work as expected if you have a split node as the first node in your flow. [(3) ]() Recommended approach from community feedback: - Lower Interruption Sensitivity so the agent doesn't cut in as easily [(2) ]() - Slightly tweak Responsiveness (don't set it too low) - Enable Pause Before Speaking to avoid cutting users off at the start - If issues persist, try a higher-context model like GPT-4o instead of 4o-mini The interruptions may be caused by the turn-taking model misinterpreting pauses or background noise as the end of the user's turn. [(4) ]() Lowering Interruption Sensitivity is likely your most effective adjustment for this specific issue.
​ To learn more, see the sources I considered: [[1] Step 1: Configure global settings](https://docs.retellai.com/build/conversation-flow/global-setting) [[2] AI Interruptions](https://discord.com/channels/1186424376300679209/1359363109822922922) [[3] Pause Before Speaking](https://discord.com/channels/1186424376300679209/1438522585485676647) [[4] Retell AI's Turn-Taking Model for Seamless Calls | Retell AI](https://www.retellai.com/blog/how-retell-ais-turn-taking-model-ensures-seamless-calls) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , what is the best practice configuration for Responsiveness and Sensitivity for clinics?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer I don't have specific "best practice" configurations for clinic use cases documented in the available sources. Here's what the documentation says about these settings: Responsiveness (value range 0-1) : [(1) ]() , [(2) ]() - Lower value = less responsive agent (waits more, responds slower) - Higher value = faster exchanges (responds when it can) - Reducing responsiveness by 0.1 adds 0.5 seconds of agent wait time - Set it lower if you want the agent to respond slower, which can be useful when talking to folks like elderly Interruption Sensitivity (value range 0-1) : [(1) ]() , [(2) ]() - Lower value = takes longer/more words for user to interrupt agent (more resilient to background speech or user interruptions) - Higher value = easier for user to interrupt agent - When set to 0, agent would never be interrupted [(3) ]() For clinic environments, consider: Based on community feedback for similar interruption issues, decreasing Responsiveness slightly and increasing Interruption Sensitivity was suggested as a starting point for unclear audio situations. [(4) ]() Since clinics may involve elderly patients or callers providing detailed medical information, you might start with: - **Responsiveness**: slightly below default (e.g., 0.8-0.9) to give patients more time - **Interruption Sensitivity**: lower (e.g., 0.4-0.6) to avoid cutting off patients mid-explanation Test and iterate based on your specific call recordings and patient demographics.
​ To learn more, see the sources I considered: [[1] Step 2: Configure the basic settings](https://docs.retellai.com/build/single-multi-prompt/configure-basic-settings) [[2] Step 1: Configure global settings](https://docs.retellai.com/build/conversation-flow/global-setting) [[3] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[4] Agent interrupting users on speakerphone calls](https://discord.com/channels/1186424376300679209/1442882951246512168) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , got it , Enable Pause Before Speaking to avoid cutting users off at the start - i have configured 0.4 sec
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer Good—0.4 seconds is a reasonable starting point for Pause Before Speaking. [(1) ]() Keep in mind this setting specifically applies to the beginning of the call when the agent speaks first—it makes the agent wait for the configured duration before speaking. , [(2) ]() If you're still experiencing interruptions throughout the conversation (not just at the start) , the key setting to adjust is **Interruption Sensitivity**—set it lower to make the agent more resilient to background speech or user interruptions. [(1) ]() , [(2) ]() , [(3) ]() You may need to experiment with the combination of: - **Pause Before Speaking**: 0.4s (your current setting) - **Interruption Sensitivity**: try lowering it (e.g., 0.6-0.8) - **Responsiveness**: slightly lower if needed, but don't go too low or call quality suffers [(4) ]() Monitor your call recordings to see if the adjustments help with the specific interruption patterns you're experiencing.
​ To learn more, see the sources I considered: [[1] Step 1: Configure global settings](https://docs.retellai.com/build/conversation-flow/global-setting) [[2] Step 2: Configure the basic settings](https://docs.retellai.com/build/single-multi-prompt/configure-basic-settings) [[3] Handle background speech & noise](https://docs.retellai.com/build/handle-background-noise) [[4] AI Interruptions](https://discord.com/channels/1186424376300679209/1359363109822922922) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy also the voice is not stable, open ai model with cimo voice. what should i do? moreover, let me know what will be the best way to speed up the call without affecting the quality
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer I don't have specific information about "Cimo voice" stability issues or detailed voice configuration troubleshooting in the available sources. Regarding voice stability: The sources mention that Text-to-Speech (TTS) technology converts the AI's textual response into natural, human-like speech. [(1) ]() Some TTS systems allow customization to reflect regional accents or preferred speaking styles. If you're experiencing unstable voice output, consider: - Trying a different voice option in your agent settings - Checking if the issue persists across different call types (web vs phone) - Reaching out to Retell support for specific voice troubleshooting Regarding speeding up calls without affecting quality: Retell AI achieves sub-600ms latency performance, which is crucial for natural conversation flow. [(2) ]() To improve call speed: - **Responsiveness setting**: Increasing responsiveness makes the agent respond faster when it can—higher value means faster exchanges [(3) ]() - However, be cautious: setting responsiveness too high may cause the agent to respond prematurely if users speak slowly, which can increase latency due to interrupted API calls A community suggestion was to add a general guideline in your prompt to "be patient and give time for the user to finish" while also modulating interruption sensitivity if that helps. For specific voice stability issues with "Cimo voice," I'd recommend contacting Retell support directly. Check out these potentially useful sources: - [Contact Sales]() - [)
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
​ To learn more, see the sources I considered: [[1] Top 5 Best AI Voice Agent Platforms | Retell AI](https://www.retellai.com/blog/best-voice-ai-agent-platforms) [[2] ](https://www.retellai.com/resources/retell-ai-receptionist-understands-calls-better) [[3] Discord message from @theognis. on 2024-08-12](https://discord.com/channels/1186424376300679209/1186462578310910112/1272623893030637643) ​
s
@AI Evy , what will be the best practice for voice speed while speaking multilingual with cimo?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer I don't have specific best practice documentation for "Cimo" voice speed settings with multilingual configurations in the available sources. Here's what I can share about voice speed settings: Voice Speed Configuration: [(1) ]() -
voice_speed
controls speed of voice with a value ranging from [0.5, 2] - Lower value means slower speech, while higher value means faster speech rate - If unset, default value 1 will apply For Multilingual Agents: - Set
language
to
multi
for multilingual support - For specific languages like Hebrew (
he-IL
) , you can set the language parameter directly Community Feedback on Non-English Languages: [(2) ]() If your agent is talking too fast between responses in non-English languages, STT might miss the start of user input. You can adjust the voice response speed in the voice model settings or prompt. Recommendation: For multilingual use with non-English languages, consider starting with a slightly lower voice speed (e.g., 0.8-0.9) to help ensure speech recognition catches user input correctly. Test and iterate based on your specific language combinations. I don't have specific documentation about "Cimo" voice configurations. You may want to reach out to Retell support for voice-specific guidance. Check out these potentially useful sources: - [Create Voice Agent]() - [Contact Sales]()
​ To learn more, see the sources I considered: [[1] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[2] TTS](https://discord.com/channels/1186424376300679209/1387743989490389052) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy what will be the best practice configuration for multinlingual agent?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
### Answer I don't have comprehensive "best practice" documentation specifically for multilingual agent configuration in the available sources. Here's what I can share: Language Setting: [(1) ]() Set the
language
parameter to
multi
for multilingual support. Available language options include: -
en-US
,
en-GB
,
en-AU
,
en-IN
,
en-NZ
-
de-DE
,
es-ES
,
es-419
,
fr-FR
,
fr-CA
-
ja-JP
,
pt-PT
,
pt-BR
,
zh-CN
,
ru-RU
,
it-IT
,
ko-KR
-
he-IL
(Hebrew) ,
ar-SA
(Arabic) - And many more regional variants Multilingual Support Update: [(2) ]() Multilingual mode now supports seamlessly switching between English, Spanish, French, German, Hindi, Russian, Portuguese, Japanese, Italian, and Dutch (using nova 3) . Key Configuration Settings to Consider: [(1) ]() , [(3) ]() - **`voice_speed`**: Value ranging from [0.5, 2]. Lower = slower speech, higher = faster. Default is 1. - **`responsiveness`**: Value ranging from [0, 1]. Lower = less responsive (waits more) . Reducing by 0.1 adds 0.5 seconds of wait time. - **`interruption_sensitivity`**: Value ranging from [0, 1]. Lower = more resilient to interruptions and background speech. - **`boosted_keywords`**: Add commonly used terms, names, or brand words to improve transcription accuracy for domain-specific vocabulary. Speech Normalization: [(1) ]() Enable
normalize_for_speech
to convert entities like dates, currency, and numbers into spoken form for more consistent speech synthesis. For specific multilingual voice configurations, I'd recommend testing different combinations and reaching out to Retell support for language-specific guidance.
Check out these potentially useful sources: - [Create Voice Agent]() - [Contact Sales]()
​ To learn more, see the sources I considered: [[1] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[2] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) [[3] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
s
@AI Evy , is there any configuration of "End of Turn Timeout"?
a
Thanks for the follow-up, @Smartb. I'll get back to you soon.
hat might help identify the right setting to adjust.
### Answer I don't have specific documentation about an "End of Turn Timeout" setting by that exact name in the available sources. However, Retell AI handles turn-taking through its proprietary turn-taking model rather than a simple timeout-based approach. [(1) ]() , [(2) ]() Related settings that influence turn detection: - Responsiveness (value range 0-1) : Controls how responsive the agent is. Lower value means the agent waits more before responding. Reducing responsiveness by 0.1 adds 0.5 seconds of agent wait time. [(3) ]() - Interruption Sensitivity (value range 0-1) : Controls how sensitive the agent is to user interruptions. Lower value means it takes longer/more words for the user to interrupt the agent. When set to 0, the agent would never be interrupted. [(4) ]() , [(5) ]() - Reminder Trigger (
reminder_trigger_ms
) : If set (in milliseconds) , will trigger a reminder to the agent to speak if the user has been silent for the specified duration after some agent speech. Default value is 10000 ms (10 seconds) . [(4) ]() , [(5) ]() Retell AI's turn-taking model uses contextual analysis combining sound signals with semantic understanding to determine whether a user is pausing or finished speaking, rather than relying solely on silence duration. [(1) ]() Could you clarify what specific behavior you're trying to achieve? T
​ To learn more, see the sources I considered: [[1] VAD vs. Turn-Taking Endpoints in Conversational AI | Retell AI](https://www.retellai.com/blog/vad-vs-turn-taking-end-point-in-conversational-ai) [[2] Retell AI's Turn-Taking Model for Seamless Calls | Retell AI](https://www.retellai.com/blog/how-retell-ais-turn-taking-model-ensures-seamless-calls) [[3] Step 1: Configure global settings](https://docs.retellai.com/build/conversation-flow/global-setting) [[4] Get Agent Versions](https://docs.retellai.com/api-references/get-agent-versions) [[5] Get Voice Agent](https://docs.retellai.com/api-references/get-agent) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.