Hi Vapi team, I am implementing text chat using directly the API (without the official widget). Currently: 1. I start a session with vapi.start(). 2. I mute the user's microphone with vapi.setMuted(true) 3. However, the agent keeps sending audio (we can't unmute it). My doubts is: - Is there a more efficient method to start text-only sessions that does not consume voice resources (WebRTC, microphone, etc.)? - Is there a way to completely mute the agent audio for a purely text-only experience? - How are credits/resources accounted for when using the API in text vs. voice mode? I am looking to implement an efficient chat experience without unnecessary audio resource consumption. Thank you!

Having trouble getting Vapi to work on my PC. I was able to open the dashboard on another laptop with my email addresses. I've restarted my browsers and PC, and checked to make sure JavaScript was enabled, but I'm still getting this grey display. its a brand new desktop, is there something with windows 11 OS or hardware that makes this happen? any insight on a fix would be greatly appreciated. https://cdn.discordapp.com/attachments/1425619132589539450/1425619132946317392/troubleshoot.jpg?ex=68e83ec8&is=68e6ed48&hm=695473cc4eeb2a33e0347b288b882ea6a7db401d90c0b54f0c57981e5d2d6c22& https://cdn.discordapp.com/attachments/1425619132589539450/1425619133319479366/troubleshoot2.jpg?ex=68e83ec8&is=68e6ed48&hm=d819896612d4956d1e93b77f20bc9b39ce742bd363543eacf427ae79dfc01dca& https://cdn.discordapp.com/attachments/1425619132589539450/1425619133667475527/Java.jpg?ex=68e83ec8&is=68e6ed48&hm=edb56cf11dbeecfea99f85cedeb4437e259558cef112c813b43d79acdd6166a5&

Hi! When I use the TTS nova-2-general-pt-br model in Vapi, it detects the information incorrectly. However, when I download the same recording from Vapi and send it to Deepgram, it transcribes the conversation correctly. Could you please explain why the same model from the same provider works in Deepgram but fails in Vapi? ssistant id:8113c4cd-c9f1-4523-847a-15be5fbf0850 17:29 calls example: 0199c575-5dd7-7bb5-8e6e-eb0dfb7ff535 17:29 0199c574-995f-7aa0-a241-6e4afd39dc58

I'm just clicking "Talk to Assistant" on the dashboard but it's giving me this error https://cdn.discordapp.com/attachments/1425567841574457516/1425567841809207379/image.png?ex=68e80f03&is=68e6bd83&hm=2ca9fc774be83d99e591c6eaaaff4b8c6359379e94472de1741ba5031b76f86d&

i tried changing the phoneCallProviderBypassEnabled to false to fix call forwarding drops but i get a HTTP 400 Bad reqeust. { "message": ["property phoneCallProviderBypassEnabled should not exist"], "error": "Bad Request", "statusCode": 400 }

Is there a pinned message or status I can check on?

My account is shared with a customer. I want to remove my payment method but can not.. Please help

Dear VAPI Support Team, We are experiencing a critical production issue with bidirectional audio transmission that is occurring frequently in our live calls. This is causing immediate business disruption as customers cannot hear our agent in real-time and the agent cannot hear customers, despite both appearing in post-call recordings. Organization ID: ac4993f3-edcf-484d-a00c-c654245a4a6d Issue Overview Symptoms: Client cannot hear agent speaking during live call Agent cannot hear client's voice during live call Post-call recordings and transcripts show complete conversation (both sides) This is happening frequently, not just isolated incidents Technical Analysis We compared a working call vs. problematic calls and found critical differences: Working Call (0199ab4d-ff1f-7668-88eb-3f9774085934): ElevenLabs: Normal voice processing Deepgram: Final transcripts contain client messages Result: Complete bidirectional audio Problematic Calls (Example: 0199ab28-821d-7337-93d2-684cca7bfa65): ElevenLabs: Agent message marked as "voice cached" Deepgram: NO final transcript of client speech Recording: Only agent voice appears, client audio missing Result: Complete audio transmission failure Additional Affected Call IDs: 0199aaf2-9f2c-7eeb-8cfe-8d1e89c1753d 0199a72b-af59-7ffe-bf98-deca3606a396 0199a6e4-b2a3-7772-9f7d-c3547d8e6038 0199ab1c-c988-722d-8f00-d5081d641fe0 0199ab1b-d323-7113-a168-0612caeab07a Key Findings ElevenLabs "voice cached" indicator appears in failing calls - suggests audio isn't streaming in real-time Deepgram shows no client audio transcription despite caller being connected - indicates audio pipeline failure Recordings contain full conversation - issue is specifically with real-time streaming, not recording System Configuration Integration: Go High Level + VAPI + n8n + Twilio Transcription: Deepgram Voice Synthesis: ElevenLabs Affected: Both inbound and outbound calls

I recently attended VapiCon, and Sean DuBois, w/ Open AI held a workshop where he configured a small ESP device to talk to a VAPI assistant. In his GitHub docs, he uses the endpoint “https://staging-webrtc.vapi.api/“ I’m wondering if it’s pointing at some kind of vapi dev environment that I’m unable to access because I keep getting an error when I attempt to do so. Is there a different endpoint URL I should be using for WebRTC? GitHub: https://github.com/VapiAI/vapicon-2025-hardware-workshop

I sent invites to team members over a week ago to join our VAPI organization, but they never received the invitation emails. We've thoroughly checked spam folders and email filters - nothing came through. This is blocking our product deployment. There's no option to resend invites in the system, and we're stuck unable to add our development team.

As I provide basic info about pricing using an .md file either as knowledge tool or as directly attached files the Assistant never retrieves any info saying there is no info in KB. No matter what i try, perhaps .md files are not supported? should i upload PDF? btw: attaching .md files directly to the assistant via files gives an error but using the Files menu and uploading them one by one does allow the files to be uploaded. (but this is another issue) What has changed? this used to work before Example When using only attached Files (to Assistant): "question": pricing? defaultQueryTool { "knowledgeBaseNames": [ "default" ] } Response Data: Success. When using only query_tool (removing the files from the assistant, relying the query_tool tool calling) {"knowledgeBaseNames": ["MyCompany_Knowledge_Base"]} Response Data: Success. ---- Here is the actual .md file used in both cases, couldn't be more simple MyCompany Pricing Plans Essential Plan: - Price: $0 - Credit Cards processing fee: 0.5%* - Features: Pay as you go, Single user, Payment links, Transaction History, Pass Off Transaction Fees Business Plan: - Price: $10 Every month - Features: All Essential features, Multiple Users, Automatic Recurring Transactions, Hold to Pay, Vanishing payment links, Priority Support Premium Plan: - Price: $35 Every month - Features: All Business features, 0%* Credit Cards processing fee Additional Information: - *Excluding gateway fees. For more information on Stripe's fees, visit Stripe.com - MyCompany offers affordable, transparent, and tailored payment solutions for businesses of all sizes, including small business owners, non-profit organizations, and growing enterprises. - No hidden fees or long-term contracts. Why Choose MyCompany? - Transparent Pricing: No surprises, no hidden fees. - Flexibility: Only pay for what you use, with the ability to pass transaction fees to your customers.

Hi Vapi, We hope this message finds you well. We're writing to follow up on the Vapi credits code you kindly provided for our upcoming congress, the 3rd Annual Meeting of the Italian Society of Artificial Intelligence in Medicine (SIIAM). This year's meeting will be held in Naples on October 10-11, 2025, with the theme "AI for Italian medicine: navigating innovation, development, and clinical practice." During a recent test, we found that the coupon is currently marked as "non-valid." As our congress is scheduled for October 10-11, 2025, we would appreciate it if you could ensure the code is functional and valid for those dates. Thank you for your attention to this matter. We look forward to your feedback. Best regards, Francesco Baglivo

I have created a tenant issue reporting system using VAPI and Make.com. Part of the system makes a transient outbound call triggerred from within make.com. The call asks a contractor if they can handle an issue (previously reported by the tenant), if "No" then the transient call makes a function call to a webhook in a separate make.com scenario which then enables the selection of the next contractor (from a google sheet). All this works well EXCEPT when the transient call encounters a voicemail or silence (no answer) - the transient call "times out" after a number of seconds and does NOT trigger the function tool (to alert the make.com scenario) so it can select the next contractor. I am seeking help on how I can ensure the transient call does not timeout before triggering the function tool. Any advice would be appreciated.

One of my agents -> that is not very often used was repetedly called maybe 100+ times within a few hours, by another agent and its drained my account. This will impact my business operations and possibly loose my first few customers ive introduced to the platform. Can someone please assist.? #cybersecurity😩

We have been using Vapi for some time now and today while deleting unused SIP credentials from the web interface, and possibly due to a web portal bug, our live credential got deleted. We are trying to recreate this but it is giving errors: { "provider": "byo-sip-trunk", "name":"VCAGENT1002", "gateways": [ { "ip": " cc04727.atmoviacc.viacloud.com ", "port": 5061, "outboundEnabled": true, "outboundProtocol": "tls/srtp" } ], "outboundAuthenticationPlan": { "authUsername": "", "authPassword": "Jd****ZH6F", "sipRegisterPlan": { "domain": "cc0****ud.com", "username": "1002", "realm": "" } }, "outboundLeadingPlusEnabled": false } The error is: { "message": "Couldn't validate SIP trunk credential. SIP gateway creation failed.", "error": "Bad Request", "statusCode": 400 } We have verified the credentials are ok, and when adding with IP instead of the FQDN in gateway 0, the credential is being added (and we are then updating the gateway 0 with api from ip to FQDN which is being accepted and updated in the credential) but the calls are not working, it seems the sip user is not being registered at all from Vapi side.

I Keep having a "Sign In Failed" popup. My Credentials are valid I tried other browsers, but I still keep getting the same popup

Here is the call id but its basically all of my most recent calls: 0199c3d2-5cd8-7aa1-b0ef-ef51f2a74350 My agent speaks the greeting message yet does not respond until the 12th second of the call which is extremely frustrating. Please help me get this fixed ASAP, I've got a client waiting.

My VAPI chat widget for my website, when opened on mobile, is too big for the screen (see attached photo). Do you know how to make it responsive to formatting on mobile? https://cdn.discordapp.com/attachments/1425479864894361670/1425479865125306451/IMG_4245.jpg?ex=68e7bd13&is=68e66b93&hm=298daa813d7c6bfdd50e4b68da55c7fa1eb2367e56e1b87ce2c918dfad172ac1&

I keep getting this issue reoccuring what are the possible ways you could provide into solving these "Couldn't Get Assistant.

assistantId

1a93fe38-c2de-4237-a130-febabe4921b9 Does Not Exist."

CallId: 0199bfa0-8398-788b-8bc1-d33d506057ec Several calls from yesterday are not usable. We have no detail in the dashboard. Blank recording, No summary, No transcript... This call lasts > 4min, we have no id of the output of this call and customer is asking details on it. Other call ids in the same situation: 0199bf9f-8c99-711e-a9dd-67dbb41751a1 0199bf9d-984d-788b-8bbb-81df8a7dbd52

We're having problems with an inbound call to an assistant with MCP tools attached. From what we were able to gather the MCP tool is treated as function. Tool discovery is not being used. What we can see in the logs is that a call was made to the endpoint with the tool name corresponding to the MCP tool name that leads to MCP error -32602.

Call Id: 0199c328-ae62-755c-bf51-6c90cef358fb Timestamp: 6min 58s Expected Response: AI Assitant should response Actual Response by AI: No response, It seems the agent crashed or the maybe from your side gpt realtime connection was lost.

Hi, I'm using the new integration of the OAuth 2.0 credentials. It seemed to work fine in the beggining but now I don't think the refresh token flow is being properly handled as I'm getting expired tokens on my webhooks/tools endpoints. Can you shime some light on this issue? Are new tokens generated per request? Are the tokens only generated once per integration? If so, how do you handle the expired ones? Also, I used to use custom headers to place my Auhtorization header but since adding this integration I can't properly delete them. I delete them, save the webhook/tool and after the refresh there they are again with the same values.. Thanks!

We want to create 66 assistants, each with its own number. We’d like to ask about concurrency — does it mean that if we have 66 assistants and 10 are included (so 66 − 10 = 56 active), we have 56 concurrent instances and we have to pay 56 * 10 $ line per month? Or does concurrency mean having all 66 calls happening at the same time? We need this information as soon as possible.

Hello, I plugged my Telnyx and my Vapi together, when I call the number and look at the logs, it says DATABASE-ERROR. I don't know how to handle that Callid: 0199c2e2-7560-7553-9d00-3c3d72c34087 https://cdn.discordapp.com/attachments/1425398411594567700/1425398411754078299/image.png?ex=68e77137&is=68e61fb7&hm=1f993541223751648054de05ee4f8b8c6196e7eab67c8acf32a42c0084b1f2ec&

I've encountered this issue with VAPI, that only incoming calls are firing webhooks, outgoing calls are not firing end-of-call webhooks. One of the outgoing calls with the issue is this one: **Call-ID: **0199c2bb-3bca-7ddf-94f6-52fb648f2feb **Time: **9:34:17 AM(+02:55.63) Additional Information: Screenshots and reproduction steps on this link or the attached screenshots here: https://app.screencast.com/nlwQuEIFtH5Zv Methods tried: - Outbound Call triggered by API (Dont work) - Outbound Call triggered by Web-Interface (Dont work) - Inbound Call (Works, but not the use-case for us) - Talk to Assistant (Works, but not the use-case for us) Already tried attempts: - Changing the webhook site url - Changed the assistant json for API call - Changed the webhook endpoint service (debugging http bin) Observations: - Call logs are created, end call event in outbound triggered by assistant works - Webhook logs is not inserting any webhook log for outbound calls by api - Webhooks for incoming calls are fired Used tools: - Vonage - Deepgram Nova - Elevenlabs Flash 2.5 - GPT 4.1 https://cdn.discordapp.com/attachments/1425397917451161662/1425397917799284857/2025-10-08_09-36-58.png?ex=68e770c2&is=68e61f42&hm=7de5ff4317a179a2dc38fa593e32f3c475827abb23c6e5128c1767e7a7c6dfb7& https://cdn.discordapp.com/attachments/1425397917451161662/1425397918348873790/2025-10-08_09-40-27.png?ex=68e770c2&is=68e61f42&hm=789223441fb1f91ede0db8a42eac447291f2e70d66449c5006e825588c416f0f& https://cdn.discordapp.com/attachments/1425397917451161662/1425397918747066418/2025-10-08_09-43-38.png?ex=68e770c2&is=68e61f42&hm=c88bcb0c9de6d4619d47e4dea05a65799b80f3820de79707730ab45a583fb658& https://cdn.discordapp.com/attachments/1425397917451161662/1425397919019962418/2025-10-08_10-10-32.png?ex=68e770c2&is=68e61f42&hm=b860d22d90a92051b8d7a3906abd01e9a5eb3280f0dcc66192d62df2889bed67& https://cdn.discordapp.com/attachments/1425397917451161662/1425397919451844789/2025-10-08_10-11-37.png?ex=68e770c2&is=68e61f42&hm=cc2b93d251bdebe9c43238b973673d524573ce1a508dea6a11ef05b617e50944& https://cdn.discordapp.com/attachments/1425397917451161662/1425397919711887451/2025-10-08_10-13-35.png?ex=68e770c2&is=68e61f42&hm=c3114a2dad1a4f2421f29a8bcc36951226c6a7fbfb0ef9fa917ddf7be681a16c&

Suggestion: Make sure your transfer destination number is in E.164 format (e.g.,

+491234567890

for Germany) and that your transfer tool includes the required

destination

property. Here’s an example configuration for your transfer function:

json
"parameters": {
  "type": "object",
  "properties": {
    "destination": {
      "type": "string",
      "enum": ["+491234567890"], // Replace with your German number
      "description": "The destination to transfer the call to."
    }
  },
  "required": ["destination"]
}

Double-check this setup in your tool—incorrect number format or missing parameters are the most common reasons for transfer failures with Twilio SIP trunks and international numbers. Sources: - [Vapi documentation: call forwarding](https://docs.vapi.ai/call-forwarding) - [Debug call forwarding drops](https://docs.vapi.ai/phone-calling/transfer-calls/debug-forwarding-drops)

Hi VAPI Support Team, I’m encountering an issue when creating Customer objects via the API. The request works successfully when I include only name and number, but it returns a 500 Internal Server Error whenever I include assistantOverrides.variableValues. orgId: 727f8c9d-9632-4e41-9268-1919aa9c97c7 Below are the details: ===================================================== ✅ This works: Request:POST https://api.vapi.ai/customer { "name": "jonny", "number": "+19174761638" } Response: { "id": "87011052-9735-41bd-93d2-5b4422b56365", "orgId": "727f8c9d-9632-4e41-9268-1919aa9c97c7", "number": "+19174761638", "createdAt": "2025-10-08T01:32:13.207Z", "updatedAt": "2025-10-08T01:32:13.207Z", "name": "jonny" } ================================================== ❌ This fails: Request:POST https://api.vapi.ai/customer { "name": "jonny", "number": "+19174761638", "assistantOverrides": { "variableValues": { "Cur_State": "NY" } } } Response: { "statusCode": 500, "message": "Internal server error. Contact support@vapi.ai for help." } ================================================= Could you please confirm what am I missing? Thank you for your help! Best regards, Ronak Shah shah.ronak@columbia.edu https://cdn.discordapp.com/attachments/1425297093270179923/1425297093643468922/image.png?ex=68e712db&is=68e5c15b&hm=7b9d884c588edec2af2f9627c607ab85d1d3b7789044dba0a3e2ef972722f57e& https://cdn.discordapp.com/attachments/1425297093270179923/1425297094184276068/image.png?ex=68e712db&is=68e5c15b&hm=456fecae4950185043c9fe095823bde68c7242390f4c4046bb597cbba4dd97cb&

Hello, I’m having a problem with voicemail detection. I’ve tried different providers — Vapi, OpenAI, and others. In about 60–70% of cases, the voicemail is detected and left correctly. However, in the remaining cases, the agent detects the voicemail but hangs up after saying just a few words. I’ve tried changing various parameters in the voicemail detection tool, but it doesn’t seem to help. Does anyone have any potential solutions or ideas? https://cdn.discordapp.com/attachments/1425294154363306085/1425294154564636792/local-file-8847793341655084184.jpg?ex=68e7101f&is=68e5be9f&hm=89e09e1f532440c67e36d79ded342c5f938b64c53a305f3de6f7657e2882ea02&

"end chat" coming by default in web widget for voice mode aswell, i dont see any way to hide it, is this a bug?