https://retellai.com/?utm_source=discord&utm_medium=seo logo
Powered by
I am in France and my voice agent is in French
# ❓・ai-evy-support
m
@AI Evy I am in France and my voice agent is in French. Whenever the caller speaks for more than 5 seconds, the agent interrupts as if the caller is done speaking. Why is that? I checked latency, lowered and increased reponsiveness. I used my voice, a pro voice, and other voices. I reduced the prompt, reduced the knowledge to a minimal amount. I even mentioned in the prompt to not interrupt until the caller is clearly done with the questions or response. It keeps doing it. When I try other's people demos it does not do that. What am I doing wrong?
a
Thanks for tagging me @Maxime. I'll look through relevant sources and get back to you shortly.
### Answer The interruption behavior you're experiencing is controlled by the 
interruption_sensitivity
 setting, not just responsiveness. [(1) ]() Key setting to adjust: 
interruption_sensitivity
controls how sensitive the agent is to user interruptions. Value ranging from [0,1]. Lower value means it will take longer / more words for user to interrupt agent, while higher value means it's easier for user to interrupt agent. If unset, default value 1 will apply. When this is set to 0, agent would never be interrupted. Recommendations: 1. Lower your 
interruption_sensitivity
 — Try setting it closer to 0 (e.g., 0.2 or 0.3) . This will require more words/longer pauses before the system thinks the caller is done speaking. 2. Adjust 
responsiveness
 — Lower values mean the agent will "wait more, respond slower." [(1) ]() [(2) ]() 3. Check your 
language
setting — Ensure it's set to 
fr-FR
for French (France) . The speech recognition language setting affects how the system processes speech patterns. [(1) ]() The combination of high 
interruption_sensitivity
(default is 1) with high 
responsiveness
(default is 1) can cause the agent to jump in too quickly, especially with longer utterances.
​ To learn more, see the sources I considered: [[1] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) [[2] Question about Responsiveness](https://discord.com/channels/1186424376300679209/1352059180164317286) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
it's the opposite, the AI voice agent interrupts the caller
@AI Evy It's the opposite, the AI voice agent interrupts the caller
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer Je comprends mieux maintenant — l'agent IA interrompt l'appelant, pas l'inverse. Le paramètre clé à ajuster est 
responsiveness
, pas 
interruption_sensitivity
. [(1) ]() - 
responsiveness
 contrôle la réactivité de l'agent. Une valeur plus basse (proche de 0) signifie que l'agent attendra plus longtemps avant de répondre, tandis qu'une valeur plus haute signifie des échanges plus rapides. Recommandations : 1. Diminuez 
responsiveness
 — Essayez une valeur autour de 0.5 ou plus basse. Réduire la réactivité de 0.1 ajoute environ 0.5 secondes de temps d'attente pour l'agent. [(2) ]() 2. Vérifiez le paramètre 
language
 — Assurez-vous qu'il est bien défini sur 
fr-FR
pour le français (France) . Ce paramètre influence la reconnaissance vocale et le traitement des pauses naturelles dans la parole. [(1) ]() 3. Problème similaire signalé — Un utilisateur a rencontré un problème similaire avec des appels en haut-parleur où l'agent interrompait les utilisateurs. La suggestion était de diminuer 
responsiveness
et d'augmenter 
interruption_sensitivity
. [(3) ]() En résumé : une 
responsiveness
trop élevée (par défaut 1) peut faire réagir l'agent trop vite, interprétant une pause naturelle comme la fin du tour de parole de l'appelant.
​ To learn more, see the sources I considered: [[1] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) [[2] Step 2: Configure the basic settings](https://docs.retellai.com/build/single-multi-prompt/configure-basic-settings) [[3] 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.
m
@AI Evy Yeah but then there is a massive silence between each reponse. Whether it's on speaker or not on speaker mode. How do other people do it?
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
​ 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] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) ​
### Answer The challenge you're facing is finding the right balance between these two settings. Here's how to approach it: The key is to adjust both settings together: - 
responsiveness
 — Controls how quickly the agent responds. Lower values add wait time (reducing by 0.1 adds ~0.5 seconds of agent wait time) . [(1) ]() - 
interruption_sensitivity
 — Controls how easily the agent gets interrupted. Lower values make the agent more resilient to background speech or user interruptions. [(2) ]() , [(1) ]() Recommended approach: 1. Start with moderate 
responsiveness
 (e.g., 0.7-0.8) rather than going too low — this reduces the silence gap while still giving callers more time to speak. [(3) ]() 2. Lower 
interruption_sensitivity
 (e.g., 0.3-0.5) — This makes the agent more resilient to pauses being interpreted as end-of-speech, without adding silence. 3. Enable 
backchanneling
 — This allows the agent to interject with acknowledgment phrases like "yeah" or "uh-huh" during longer user utterances, which creates more natural conversation flow without full interruptions. [(2) ]() , [(3) ]() The combination that often works: moderate responsiveness + lower interruption sensitivity + backchanneling enabled. This lets the agent acknowledge the caller is speaking without fully interrupting, while still responding reasonably quickly when they're done.
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
@AI Evy Yeah it's not working. It keeps having the same issue
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
2. Try 
stt_mode
setting You can try switching between 
fast
and accuracy-focused speech-to-text modes, as this affects how speech patterns are processed. [(3) ]() 3. Check 
denoising_mode
 Try setting 
denoising_mode
to 
noise-cancellation
if you haven't already — this may help with audio clarity issues that could cause false end-of-speech detection. 4. Consider reaching out to support A similar speakerphone interruption issue was escalated to Retell support, where they analyzed specific call IDs to diagnose the problem. [(4) ]() You may want to provide a call ID to the Retell team for deeper investigation. If none of these work, I'd recommend reaching out via [Contact Sales]() or checking [Retell AI GitHub]() for additional support options.
### Answer ### Answer Since adjusting 
responsiveness
and 
interruption_sensitivity
hasn't resolved the issue, here are additional approaches from the community and documentation: 1. Prompt-based interruption handling Some users have had success adding explicit interruption handling instructions to their prompt: [(1) ]() , [(2) ]()
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.
​ To learn more, see the sources I considered: [[1] Discord message from @sanjay_69738 on 2025-01-07](https://discord.com/channels/1186424376300679209/1294891368505937952/1326287312744615986) [[2] Discord message from @bajacheebo on 2025-01-06](https://discord.com/channels/1186424376300679209/1294891368505937952/1325906975183081564) [[3] Update Voice Agent](https://docs.retellai.com/api-references/update-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.
m
It's not the caller interrupting the AI. That's fine. I need help with the AI agent interrupting the caller. What prompting should I add for this?
@AI Evy It's not the caller interrupting the AI. That's fine. I need help with the AI agent interrupting the caller. What prompting should I add for this?
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer --- The primary control for preventing the AI agent from interrupting the caller is the 
responsiveness
 setting, not prompting. [(1) ]() Key setting: - 
responsiveness
 — Value ranging from [0,1]. Lower value means the agent will "wait more, respond slower," while higher value means faster exchanges. Default is 1. Prompt-based approach: While settings are the main lever, you can add style guardrails to your prompt. The recommended structure includes: [(2) ]()
Copy code
## Style Guardrails
Be concise: Keep responses under 2 sentences unless explaining complex topics.
Be conversational: Use natural language, contractions, and acknowledge what the caller says.
Be empathetic: Show understanding for the caller's situation.

## Response Guidelines
Ask one question at a time: Avoid overwhelming the caller with multiple questions.
Confirm understanding: Paraphrase important information back to the caller.
Recommendation: 1. Lower 
responsiveness
 significantly (try 0.3-0.5) — this is the primary control for making the agent wait longer before speaking [(1) ]() 2. Add a style guardrail in your prompt like: "Wait for the caller to fully complete their thought before responding" 3. Enable backchanneling (
enable_backchannel: true
) — this lets the agent use acknowledgment phrases like "yeah" or "uh-huh" during longer utterances instead of fully responding The issue is likely that 
responsiveness
is too high, causing the agent to interpret natural pauses as the end of the caller's turn.
​ To learn more, see the sources I considered: [[1] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) [[2] Prompt Engineering Guide](https://docs.retellai.com/build/prompt-engineering-guide) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
Appreciate the feedback! Feel free to ask me again with additional guidance or context.
m
@AI Evy I added " If you think user is still thinking or hasn't finished speaking, output exactly the following: NO_RESPONSE_NEEDED" to my prompt so that the ai agent doesn't say anything when the person is not done talking but now the ai agent literally says no response needed when the caller is thinking. How can I resolve this?
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer This is a known issue where the LLM outputs "NO_RESPONSE_NEEDED" as spoken text instead of treating it as a silent command. [(1) ]() Two solutions: 1. Try switching to GPT-4o — One community member reported that changing the model to GPT-4o resolved this issue. [(2) ]() 2. Make your prompt more explicit — Update your instructions to be clearer about the exact output format: [(1) ]()
Copy code
## Silence Rules
- If the user says "hold on", "give me a second", "wait", or similar, respond with ONLY this exact text with nothing before or after: NO_RESPONSE_NEEDED
- If you detect the user is still thinking, pausing, or not ready to respond, output ONLY: NO_RESPONSE_NEEDED
- When you output NO_RESPONSE_NEEDED, do not add ANY other text, punctuation, or explanation before or after it.
- NO_RESPONSE_NEEDED must be the complete and only content of your response.
Important note: 
NO_RESPONSE_NEEDED
stops the agent entirely — it's a terminal command. Reminder messages won't resume unless the conversation remains active. [(3) ]() If you're using a Conversation Flow agent rather than a Single/Multi Prompt agent, this issue has been reported to behave differently — the flow nodes may not fully honor the "no response" behavior and output the debug value instead. [(1) ]() , [(7) ]()
​ To learn more, see the sources I considered: [[1] no_response_needed](https://discord.com/channels/1186424376300679209/1433483896179261480) [[2] NO_RESPONSE_NEEDED ERRORS](https://discord.com/channels/1186424376300679209/1376157747783143434) [[3] NO_RESPONSE_NEEDED "disables" the Reminder Message Frequency setting](https://discord.com/channels/1186424376300679209/1407515064260624525) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
@AI Evy Which LLM model do you recommend using? My agent needs to be speaking in french.
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer --- Pour un agent vocal en français, voici les options LLM disponibles sur Retell AI : Options recommandées : - GPT-4o — Haute intelligence, bonne performance pour les conversations nuancées. Prix : $0.050/min [(1) ]() [(2) ]() - GPT-4.1 — Plus intelligent que GPT-4o, mais latence plus élevée. Prix : $0.045/min [(2) ]() - GPT-4o mini — Bon équilibre latence/prix. Prix : $0.006/min - GPT-4.1 mini — Prix : $0.016/min - GPT-4.1 nano — Latence la plus basse, prix le plus bas ($0.004/min) , mais intelligence réduite - Claude 3.5 Haiku — Prix : $0.02/min - Claude 3.5 Sonnet — Version premium pour tâches complexes. Prix : $0.06/min [(1) ]() Comparaison rapide : | Critère | Recommandation | |---------|----------------| | Intelligence | GPT-4.1 > GPT-4o > Claude 3.5 Sonnet [(1) ]() [(2) ]() | | *Latence (plus basse = plus rapide) * | GPT-4.1 nano < GPT-4.1 mini < GPT-4o mini [(2) ]() | | Prix | GPT-4.1 nano ($0.004) est le moins cher | Pour le français spécifiquement, Retell AI supporte le français avec le support multilingue (utilisant nova 3) . [(2) ]() [(3) ]() Recommandation : Pour résoudre ton problème d'interruption avec un agent français, essaie GPT-4o — un utilisateur de la communauté a signalé que passer à GPT-4o a résolu le problème de 
NO_RESPONSE_NEEDED
qui était prononcé à voix haute.
​ To learn more, see the sources I considered: [[1] Choosing the Best LLM for Your AI Voice Agents | Retell AI](https://www.retellai.com/blog/choosing-the-best-llm-for-your-voice-ai-agents) [[2] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) [[3] ](https://www.retellai.com/languages/french) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
I only have 4o real time for .5 cent and 4o real time for .12 cents
@AI Evy I only have 4o real time for .5 cent and 4o real time for .12 cents
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer The two options you're seeing are: - GPT-4o Realtime at $0.50/min [(1) ]() - GPT-4o Mini Realtime at $0.125/min These are the Realtime models that use OpenAI's native voice-to-voice API. [(2) ]() For your French agent with the interruption issue, GPT-4o Realtime ($0.50/min) would be the higher-intelligence option between those two. [(1) ]() If you need access to the non-Realtime LLM options (like GPT-4o at $0.05/min, GPT-4.1, Claude models, etc.) , those should appear when you're using a Single Prompt or Multi-Prompt agent type rather than the Realtime configuration. The available models depend on your agent setup. Check out these potentially useful sources for more details on available models and pricing: - [Platform Changelogs | Retell AI]() - [Contact Sales]() - [Retell AI GitHub]() Would you like a general answer about the differences between Realtime and non-Realtime model options?
​ To learn more, see the sources I considered: [[1] Important Pricing Adjustments, New Voice Models, and More!](https://www.retellai.com/changelog/important-pricing-adjustments-new-voice-models-and-more) [[2] 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.
m
@AI Evy What about the 5+ models with open ai? Why are they not good?
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer --- I don't have information in my sources that indicates the OpenAI models are "not good." Based on the available documentation, here's what I can share about OpenAI models on Retell AI: OpenAI GPT-4.1 family is available: - GPT-4.1 — Higher intelligence than GPT-4o, but higher latency. Price: $0.045/min [(1) ]() - GPT-4.1 mini — Price: $0.016/min - GPT-4.1 nano — Lowest latency, lowest price ($0.004/min) - GPT-4o — Price: $0.050/min - GPT-4o mini — Price: $0.006/min Intelligence ranking: GPT-4.1 > GPT-4o mini > GPT-4o > GPT-4.1 mini > GPT-4.1 nano Latency ranking (lower is faster) : GPT-4.1 nano < GPT-4.1 mini < GPT-4o mini < GPT-4o < GPT-4.1 Note on Realtime models: If you're using OpenAI Realtime models, you're limited to only OpenAI voices (you cannot use ElevenLabs or PlayHT voices) . [(2) ]() If you're only seeing Realtime options ($0.50/min and $0.125/min) , you may need to switch your agent type to access the standard LLM options listed above.
​ To learn more, see the sources I considered: [[1] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) [[2] Discord message from @pedropinto123_50933 on 2025-02-17](https://discord.com/channels/1186424376300679209/1294891368505937952/1341077024243974306) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
@AI Evy When I add check for availaibity on cal.com for the retell connection the API works but not for booking. Why
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer --- This is a known issue that several community members have reported — check availability works but booking fails. [(1) ]() [(2) ]() [(3) ]() Common causes and solutions: 1. Date/time parsing issue — The booking function may not be receiving properly formatted date/time arguments. The error 
"Date time reference does not exist, ask user to explicitly say a date time"
indicates the LLM isn't passing the datetime correctly. [(1) ]() 2. Cal.com user assignment — Verify that a testing user is added to the event type ID in Cal.com. [(2) ]() 3. Timezone mismatch — Ensure the timezone in your Retell function matches your Cal.com event settings. 4. Add current time to your prompt — Include a current time variable so the LLM is aware of the current time when booking. 5. Location field formatting — One user resolved this by adding explicit prompting to pass 
location.type
and 
location.integration
correctly in the API structure. [(3) ]() Prompt update recommendation: Add explicit instructions in your prompt for when to invoke the booking function: [(4) ]()
Copy code
When user selected a slot, please book the appointment by calling the `book_appointment` function.
The booking integration can behave inconsistently — some users report it works sometimes and fails other times. [(1) ]() [(10) ]()
​ To learn more, see the sources I considered: [[1] Discord message from @qcumbers1_cc on 2025-03-19](https://discord.com/channels/1186424376300679209/1294891368505937952/1352008218754158713) [[2] API to Cal.com keeps failing spastically.](https://discord.com/channels/1186424376300679209/1400179427480174592) [[3] Cal.com booking errors](https://discord.com/channels/1186424376300679209/1377380250954105064) [[4] Book Calendar](https://docs.retellai.com/build/book-calendar) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
@AI Evy It's saying error validating cal.com tools
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer --- The "error validating Cal.com tools" message is a known issue that has been reported by multiple users. [(1) ]() Common causes: 1. Invalid or expired API key — Verify your Cal.com API key is not expired. 2. Event Type ID not found — The error 
"Event type with id X not found"
indicates the event type may have been removed or the ID is incorrect. Confirm the event type still exists in your Cal.com dashboard. 3. Round Robin events — There have been reported issues specifically with round robin team calendar event types. Switching to a normal (non-round robin) event type has resolved this for some users. To troubleshoot: 1. Go to your Cal.com dashboard and verify the Event Type ID exists (check the URL when viewing the event type — the number at the end is your ID) . [(2) ]() [(3) ]() 2. Regenerate your API key in Cal.com under Settings → Developer → API Keys. 3. If using a round robin event, try switching to a standard event type to test. [(1) ]() 4. Create a separate support ticket with your agent ID so the Retell team can investigate further.
​ To learn more, see the sources I considered: [[1] Cal.com error validating tools](https://discord.com/channels/1186424376300679209/1432750783580999862) [[2] Book Calendar](https://docs.retellai.com/build/book-calendar) [[3] Check Calendar Availability](https://docs.retellai.com/build/check-availability) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
What's a round robin?
@AI Evy What's a round robin?
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer --- A round robin in the context of Cal.com is a team scheduling feature where appointments are automatically distributed among multiple team members. [(1) ]() [(2) ]() Based on community discussions, round robin scheduling works with "load balancing" — meaning appointments rotate between available team members rather than going to a single person's calendar. Example scenario: If you have 3 closers at your company, each with their own calendar, you can set up a round robin event type in Cal.com so that incoming appointment bookings are automatically distributed among them. [(2) ]() Note: There have been reported issues with Retell AI's native Cal.com integration when using round robin event types. Some users have found that switching to a normal (non-round robin) event type resolves validation errors. [(5) ]()
​ To learn more, see the sources I considered: [[1] Issue with book_appointment (Cal.com Native Integration) for Round Robin & Custom Location Event Typ](https://discord.com/channels/1186424376300679209/1379631392312266893) [[2] Discord message from @67d3 on 2025-03-23](https://discord.com/channels/1186424376300679209/1294891368505937952/1353500621118115930) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
I double checked all of these. Everything I did is correct. It's still not working
@AI Evy I double checked all of these. Everything I did is correct. It's still not working
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer --- Since you've verified your Event Type ID and API key are correct, and you're still getting "error validating Cal.com tools," here are additional steps: 1. Confirm you're filling in ALL required fields When adding the book calendar function, ensure you enter: [(1) ]() - Tool name (must be unique within that agent) - API Key from Cal.com - Event Type ID A similar issue was reported where the user hadn't filled in the Event Type ID field — even though check availability worked without it, the booking function requires it. [(2) ]() 2. Check for 403 vs 400 errors - 403 error = forbidden API access, meaning wrong API key or insufficient permissions [(3) ]() - 400 error = bad request, often caused by missing Event Type ID [(2) ]() 3. Known Cal.com integration issues Community members have reported ongoing issues with the Cal.com integration that appear to be on the Cal.com/Retell side rather than user configuration. [(4) ]() One user confirmed with Cal.com support that it's a known issue they're working on. Recommendation: Since you've verified everything is correct, I'd suggest reaching out directly to Retell support with your agent ID and the specific error message. You can contact them via: - [Contact Sales]() - [Retell AI GitHub]() Would you like me to provide a general answer about alternative approaches to calendar booking?
​ To learn more, see the sources I considered: [[1] Book Calendar](https://docs.retellai.com/build/book-calendar) [[2] Status Code 400 add cal api key](https://discord.com/channels/1186424376300679209/1300469686860910704) [[3] CAL.Com](https://discord.com/channels/1186424376300679209/1408212404482871469) [[4] Cal.com error validating tools](https://discord.com/channels/1186424376300679209/1432750783580999862) ​
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
@AI Evy what do you mean must be unique to that agent? (tool name)
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
### Answer --- The "tool name must be unique to that agent" means you cannot have two tools with the same name assigned to a single agent. Each tool within an agent needs a distinct name so the LLM can differentiate which tool to call. For example, if you add a 
check_availability
tool and a 
book_appointment
tool to the same agent, those names must be different from each other. You cannot have two tools both named 
book_appointment
on the same agent. I don't have explicit documentation on this specific validation rule in my available sources. If you're still encountering the "error validating Cal.com tools" message after confirming unique tool names, I'd recommend reaching out to Retell support with your agent ID. Check out these potentially useful sources: - [Book Calendar]() - [Contact Sales]() - [Retell AI GitHub]()
message has been deleted
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
m
Yeah nothing is working, who do I contact?
a
Thanks for the follow-up, @Maxime. I'll get back to you soon.
m
@AI Evy Yeah nothing is working, who do I contact?
a
Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.
message has been deleted
### Answer For support with your Cal.com integration issue, you can contact Retell AI through: - [Contact Sales]() - [Retell AI GitHub]() Since you've already verified your API key, Event Type ID, and configuration are correct, opening a support ticket with your agent ID will help the team investigate the specific "error validating Cal.com tools" issue you're experiencing.
PreviousNext
Powered by