I've got a campaign that does not begin, it stays stuck in scheduled. The details are: - ID = 8425635f-8fd1-4790-94e2-5155120a26c8 - earliestAt = Dec 19, 2025, 8:03 PM (GMT) - latestAt = Dec 19, 2025, 9:03 PM (GMT) It was created via API and it has been an hour now. There are no webhook logs, error logs, or anything else that would indicate why it is broken? Is there an issue here? In general, campaigns seem to be a particular black-box feature of Vapi - they seem to often not work for hidden reasons - it would be great if we could get some additional visibility/logging on campaigns.

Hello @Vapi Support Bot I have a conurrency limit error on my vapi dashboard. Im tryin to test new assistants that im building but it wont let me due to this error. Org id: 759d95c2-8e4e-4f48-847b-8334798bd95a

Hi everyone, I was wondering if there is currently a way for me to use Elevenlabs v3 on Vapi, it has been out for a whie but there isn't a selection for it on the UI yet, it is amazingly realistic and I really hope I can switch my clients' agents to it.

ive set up a tool called inquiry summary that ive linked the make webhook to. i;ve also set up a structured output called call summary. #vapitomake #1211483291191083018

i built an ai voice agent using vapi and i have connected it with a twilio number. But the number sometimes works fine. The voice agent talks to me but sometimes it says nothing. how to solve it? 60% of the times the number ans the call but ai doesn’t speaks. And 40% of times it works perfectly. I have also tried to solve it with free vapi numbers. But the problem didn't solve. How can i solve this call silence issue.

Hi, my assistant's chat responses are not showing in the widget on the website. They are showing in the chat logs. Here's a chat ID "548c30c1-5e4e-4017-b280-2dac6b0b831b". Here's a chat Details "8aad6806-d1ad-4771-8eaa-917c14bfbffa" #1211483291191083018

Hi Vapi Support Team, I'm following up on my previous ticket regarding the GoHighLevel (GHL) integration failure. After investigating with browser developer tools, I've identified several critical technical issues preventing the OAuth flow from completing: Issues Identified: CORS Blocking (Primary Issue) Multiple requests are being blocked by Cross-Origin Resource Blocking The OAuth callback from GHL cannot complete due to CORS policy restrictions This prevents the integration from finalizing after user authorization Content Security Policy Violations CSP is blocking script-src directive (status: blocked) Evaluation of JavaScript strings is being prevented These restrictions are interfering with the OAuth authentication flow API Communication Failures Repeated net::ERR_QUIC_PROTOCOL_ERROR 200 (OK) errors on https://api.vapi.ai/events These errors occur during the integration process Suggests backend communication issues during OAuth callback Current Behavior: Click "Connect" → Opens GHL permissions page ✓ Accept permissions ✓ Choose subaccount ✓ Redirect back to Vapi → Integration shows as NOT connected ✗ Important Note: This is not an isolated issue - multiple users are experiencing the same problem with the GHL integration, which suggests this is an infrastructure/configuration issue on Vapi's side rather than a user-specific problem. Request: Please review and update: CORS policy to allow GHL OAuth callback domains CSP directives to support the OAuth flow QUIC protocol handling for the /events endpoint I've attached screenshots of the browser console errors for your engineering team's reference. Please prioritize this as it's blocking a critical integration for multiple users. Let me know if you need any additional diagnostic information. Thank you for your assistance. https://cdn.discordapp.com/attachments/1451560639981031424/1451560640744525864/preview.webp?ex=69469eb1&is=69454d31&hm=da3202fb4074aecabe504040942d4998873d34794aa87c790b4a50c8e088bc68& https://cdn.discordapp.com/attachments/1451560639981031424/1451560641415352424/preview_1.webp?ex=69469eb1&is=69454d31&hm=caa31ee697bd69e42b1c2f16a12e91c4eb86113be14b900b5ef0d0bddb3d1629& https://cdn.discordapp.com/attachments/1451560639981031424/1451560641881051289/preview_2.webp?ex=69469eb1&is=69454d31&hm=2ad458e357628826221bdfe1ad66b4c168546806257eb1f3a1f4fcf2b5262e2f& https://cdn.discordapp.com/attachments/1451560639981031424/1451560642728165477/preview_3.webp?ex=69469eb1&is=69454d31&hm=bb7c12762fda8f3eb8d20235302afe62d0f3ddd973eb264da53a146cd0086b3d&

Hi, My company recently purchased a Twilio number, and I’m unable to import it to our Vapi account. I’m getting an error saying the number is already in use by another org. Please help fix this since this is a critical blocker. I don’t want to publish the number here. I can DM it instead. https://cdn.discordapp.com/attachments/1451513022823661658/1451513023473647718/IMG_7352.png?ex=69467258&is=694520d8&hm=c8b6a542ed90da9f9857b0b01f794a113a00853f4bf74dc429478cb4104d3106&

Is there some issues with structured outputs at the moment? I saw that some things were becoming deprecated and should be replaced by structured outputs. But I'm getting 500 errors when trying to create them through the API. Visiting the docs strangely seem to append an ?error=true flag at the end of the url which only happens to structured output related things afaik. https://docs.vapi.ai/api-reference/structured-outputs/structured-output-controller-create?error=true

Hi team, We’re facing an issue with our Vapi assistant where we’re seeing noticeable response delays during live calls. What’s happening: • The assistant sometimes pauses for several seconds before responding. • In some cases, it stays silent and then follows up with “Are you still there?”, which breaks the conversation flow. • This happens even after simple user responses like confirmations or yes/no answers. Current setup: • Prompt behavior is working as expected. • Temperature is set to 0.15 (tested lower values as well). • The issue seems to be related to latency, not response accuracy. What we’ve tried: • Prompt refinements and restructuring • Testing lower temperature values • Reviewing call logs and workflow logic • General debugging on our side Despite this, the delayed responses still persist in real-time calls. Could you please advise if there are any known causes, configurations, or best practices to reduce response latency during live calls? Happy to share more details if needed. Thanks!

Hi, I've been attempting to provision phone numbers with Vapi as the provider via the API. I noticed that the operation shows as successful in the API response, but the number is not created (shows as 'Untitled Phone Number' in the dashboard list). https://cdn.discordapp.com/attachments/1451344831354765406/1451344832441225316/Screenshot_2025-12-18_at_2.43.46_PM.png?ex=6945d5b4&is=69448434&hm=8f776658b447a176382788dc648a036d6b08e2cf3e3bc87abc91e407be1cbe8d&

I am having issues with Test Suite runs, here are the details Run 12/18/2025, 12:23:23 PM (id: 791007e7-fb61-47d3-93ff-d3e28190320e) Test Run Error Whoops, there was an error running the test suite. Please try again or reach out to us. Can you assist please as this is blocking me at the moment from testing prompt and agent

Hello! I am trying to create a simple assistant that will ONLY perform the following tasks: * read out a short, pre-defined greeting * allow the user to speak to leave their message * (optional) end the call once the user has stopped leaving their message I want to do this so that users can leave "voicemails" and I can leverage the VAPI call recording features. Problem is that I can't create an assistant without selecting an LLM and no matter how I prompt the LLM I can't seem to keep it from responding to the user and interrupting their "voicemail". Thanks in advance for your help here!

Are the vapi agents aware of which mode the user is initiating the session from, i.e call vs chat? Because for us, depedning on the mode we'd want to communicate a bit differently. Wondering if just updating the system prompt with instructions on each mode will suffice?

Native GoHighLevel tools are failing with: “GoHighLevelMCPCredential or its authenticationSession not found”. We are not using MCP tools. Only using: gohighlevel.calendar.availability.check gohighlevel.calendar.event.create get_contact / create_contact GoHighLevel integration is connected (green check). GoHighLevel MCP is not connected. This looks like provider routing incorrectly requiring MCP credential for native tools. Please check tool provider mapping for this workspace and rebind native tools to non-MCP credential path. https://cdn.discordapp.com/attachments/1451261180437073951/1451261180906573910/Screenshot_2025-12-18_211135.png?ex=694587cc&is=6944364c&hm=30904e307d81cbe5d08ef9dc32162d74bf42b9f133e0494f8b28fa2064ec5c83&

Do you have an eta for when this is gonna get fixed?

Is there any possibility to use the widget with squad instead of a single assistant?

Hi @Vapi Team, We are seeing lot of errors with this description, but the RCA is neither in the codebase nor the issue is resolved. Please help in fixing this issue soon. Latest 5 Call IDs for reference: Call ID: 019b30aa-7f3d-7334-a4f2-83acc106322b Call ID: 019b30a4-7e15-7663-9000-fd8b0ae01e3b Call ID: 019b2ff2-ec8f-7117-b299-fcdb4d97b74f Call ID: 019b2efe-f896-7338-a4fa-15078595ce73 Call ID: 019b2ee9-a30d-7cc1-a978-322326a8b01a Thanks Rajesh B K CC: @Zack Siegel

I am seeing it into all my organizations, I am also seeing it on my client's organization. https://cdn.discordapp.com/attachments/1451205554654412894/1451205554973315215/Screenshot_2025-12-18_193033.png?ex=694553fe&is=6944027e&hm=7e73970c5a08585764faa2f6b2358b2472a56861bfdc33f621e269665e4550a4&

Hi everyone 👋 I’m currently building an AI phone assistant for HR / candidate screening interviews and wanted to ask for your experience and recommendations. What do you think is the best voice model for a phone-based interview assistant in German? Key requirements for this use case are: • natural and professional tone • neutral but empathetic delivery • good German pronunciation • low latency for live conversations • stability over ~10–15 minute calls If you’ve tested different voice models (e.g. ElevenLabs, Cartesia, Azure, Rime, Vapi voices), I’d love to hear: • which model you chose • why you chose it • any trade-offs you noticed (latency vs. naturalness, etc.) Thanks a lot in advance! 🙏

I'm receiving consistent 504 Errors when creating outbound calling campaigns where the number of contacts is large. It's no problem when creating test campaigns with 1-2 recipients but as soon as the list of recipients grow to 50+ I receive this error. Two campaigns that I know were affected by this: e70122ea-fc8b-4357-beae-6a59a4fc6311 99331f3f-2c87-4012-b140-e04687c5b300 This caused the Node Vapi SDK to waste lots of my credits and irritated customers because of it's builtin retry function where it created duplicate campaigns due to this 504 error. (I've now set maxRetries to 0 because of this). Here are some campaigns that I had duplication issues with. Campaign 1: 2c0d5168-aa86-4dd4-9e59-dad001fe2cd7 82a43e29-927f-4c01-afb7-6ca0940c645b 33dc1865-027d-4594-a2a3-e9ef15a08a32 Campaign 2: 5b7cff34-fc64-4534-b6fd-0e1c64b51be7 73b0e4e4-6eb5-4d16-b2f0-257b416b3e56 c6cc4602-ccf4-4c0f-a6c0-062e5cc09950 Campaign 3: c29ff415-801d-4f2c-bc30-5130b9782be7 4f3fe393-7154-41d3-8382-e90637247ee7 114e7ae1-da17-49d6-ae80-e440b4a29fd4 Looking forward to see if you can identify what's gone wrong here. Hopefully we can resolve it together 🤞

it is spelling out names instead of pronouncing names, what is a possible fix ?

https://docs.vapi.ai/call-forwarding when will be warm transfer avaliable for BYO SIP trunk Note: Warm transfer functionality is currently available only with Twilio-based telephony systems. #SIPTRUNK #twilio

We're running an outbound calling campaign and experiencing consistent voicemail detection failures with a specific pattern. The problem: Carrier-based voicemail systems that read the phone number aloud (e.g., "4 5 6 5 3 1 8 8 7 is not available") are not being detected as voicemail. The transcriber picks this up as user speech, and our assistant starts conversing with the voicemail system. Example call ID: 019b2e8b-31a2-755f-91f0-9812c0c41719 What we've tried: All voicemail detection providers: Vapi, Twilio, Google, OpenAI — none catch this Twilio voicemailDetectionTypes settings — caused worse issues (random hangups) Voicemail tool with prompt instructions to detect number sequences — LLM can't process fast enough before responding Structured system prompts with explicit voicemail indicators Current wait Seconds is 2.5s, but increasing further creates awkward pauses for real humans Current config: Assistant ID: 7b9c01f4-2870-42c8-a161-00ecdc293461 Voicemail detection: provider: "vapi" voicemailMessage is configured and works when detection triggers The ask: Is there a way to improve detection for this specific pattern (phone numbers being read aloud by carrier voicemail)? Or is this a known gap in the detection models?

Hey team, Since yesterday (December 16, 2025), the Code Tool option has completely vanished from the dashboard when creating or editing assistants. This is happening across ALL my accounts (free and paid) and I've confirmed it's also gone for other creators (including US-based ones I've spoken to). - No announcement in changelog, blog, or anywhere else - Documentation still describes Code Tools as active and available - I've already posted about this in the community forum (no response) - Sent email to support more than 48 hours ago (no reply) - Multiple messages here in #1211483291191083018 since yesterday (still no response) This is breaking production calls that rely on serverless TypeScript execution. We need an official response ASAP: Is the Code Tool deprecated/removed? Is this a bug? When will it be back (if temporary)? Please acknowledge and clarify. Thanks.

Hey VAPI Support, Just wondering if it is at all possible to turn back and restore tools that were deleted in my clients account? Client hired another Contractor and they went in and deleted tools I had set up with their assistants in there. Safe to say these were BUSINESS CRITICAL. Anyways, I still see the assistants have the tools that were deleted's ids when in the assistant's tool assigning UI. Let me know if those would be of use! OR if this is possible, I can give you a time when these tools were still available?

We are seeing inconsistent behavior with a Vapi custom tool. What we observe: Vapi call logs show Model called tool with populated arguments (email, voice_sdr_outcome, voice_sdr_summary, booking fields). The webhook endpoint receives the tool call envelope, but function.arguments is sometimes {}. This happens intermittently. Other calls send the full arguments as expected. The tool is definitely invoked, and the server URL is correct. The workflow is active and using the production webhook URL. Resulting issue: Downstream workflows do not receive voice_sdr_* fields even though the model generated them. Questions: Under what conditions does Vapi send {} as function.arguments in the webhook? Is {} expected behavior when tool validation fails? Does the webhook payload reflect post-validation arguments rather than what the model generated? Is toolCalls[0].id the only reliable tool call identifier (vs passing toolCallId as a parameter)? We are trying to confirm whether this behavior is expected and how to configure the tool schema to prevent argument loss.

I can not find a way to block calls i got 25k spam calls i just need a solution this is really fursrated

The bot's not considering business hours and applying the following logic properly: * If inside business hours - Office Open = Live Transfer * If outside business hours - Office Closed = Schedule callback Even when I apply hard restrains on this, its still offering live transfers outside of Business Hours. I have many examples but check this: * Call ID: 019b2d1a-7b9e-7887-a7f6-5e4c6f6595b3 * Call Link: https://dashboard.vapi.ai/calls/019b2d1a-7b9e-7887-a7f6-5e4c6f6595b3 * Expected behavior: because its outside business hours, I need it to not allow Live Transfers and offer to schedule a later moment for that.

we are currently experiencing an outage with vapi with all calls throwing Ended: Call.in progress.error vapifault worker not available