Hi, I am in the UAE, and twilio and other providers don’t support UAE phone numbers. And I can’t get an SIP trunk in my case. Is there a way I can connect my phone number to VAPI using a GSM gateway?

Getting this transport error. need help https://cdn.discordapp.com/attachments/1442407847647248384/1442407847936790648/image.png?ex=6925527d&is=692400fd&hm=fc52d046b753397b8b31262d8798d410c284960fe36608789e9cd94838c8e14f&

Hello all. I am trying to configure a Custom Tool using the API. There is a property messages in the tool format. The documentation is worthless. ChatGPT's attempt to format failed. Claudes format failed. The VAPI Documentation agent gave me three different formats, all three failed. Can someone definitively tell me the format of this property? I want my agent to speak a line while my API Request is in-flight. Second part of the question: Where am I supposed to go to figure this out and any of the dozens of other questions. The documentation has nothing. The VAPI Documentation AI gave me three wrong answers. One of the promising links in the documentation took me to a github repo that had NOTHING to do with this.Whatis the right way to not be pestering VAPI support with basic API questions? -- Michael.

I'm currently experiencing an issue with the API request tool where it's copying the field for whatever request body variable I position at the top, to every other variable. The picture I attached shows what I'm trying to explain, whatever request body (e.g location, contactId, etc) I podition at the top in the request body section, that's the one that gets cloned to all others Anyone else had this same problem? Any help is appreciated https://cdn.discordapp.com/attachments/1442192072223232123/1442192072613429349/image.png?ex=69248988&is=69233808&hm=4cacf4f4cf699da48cf933a71acba00fd2dce9d0d2013e5b308318d1dd8eb6da&

I created a new squad with 2 very simple assistants to test the routing (a few lines of prompt, no more). I add an existing assistant in the squad (31 lines of prompts, nothing crazy). I don't even connect it to another agent: when I start the conversation through a web call, the agents speak extremely slowly, I can't even understand what it says. The STT when I speak is also slow and completely wrong ("I would like to know more about your services" transcribed as "It's a fun event." ??!!). I tried by removing or addind agents, creating a new squad, refreshing the browser, the problem remains. As soon as I have more than 2 agents in the squad, the problem occurs. I'm using: Chrome 142.0.7444.176 Windows 11

I'm using Vapi Squads with transient assistants (full configuration passed inline, not stored). My tool endpoints need to identify which team/account the call belongs to. My question: 1. Can I manually set an id field on a transient assistant definition that will be passed through to tool calls? 2. If not, what's the recommended way to map transient assistants to a team/account when they call tools? 3. Should I use metadata, the assistant name, or a different approach? Context: - I have 3 transient assistants in one squad (router + 2 specialists) - All need to access the same team's data when calling tools

Hey Everyone, Need some help with a SIP trunk issue that's driving me crazy So I've got my SIP trunk set up, followed the docs, whitelisted all the IPs you guys listed, tried different protocols and ports... but inbound calls just aren't working at all The weird part is: When I connect a phone number to the trunk (even a wrong one), Vapi says it's configured but doesn't throw any errors My SIP trunk backend shows nothing is actually connected to it When someone calls the number... nothing happens, no routing at all I've been troubleshooting this for a while now and can't figure out what I'm missing. Anyone know what might be going on or what I should check next? Happy to share logs/config details if needed

“Hey! I’m trying to trigger the POST https://api.vapi.ai/call endpoint from a backend webhook (GoHighLevel). What exact API key should I use for the Authorization: Bearer header when calling the REST API? Specifically: • Should I use a Private API Key or a Public API Key? • If a private key is required, where exactly in the dashboard is that key located now? • The docs show using a private bearer token, but my dashboard only displays Public API Keys under Client SDK — not a private one. • Can you point me to the correct location or regenerate the proper Bearer API key needed for server-side requests?”

Hi there, Thank you for your message. Our team is currently out of the office. We operate Monday through Friday, from 9:00 AM to 8:00 PM Pacific Standard Time (PST). We’ll get back to you as soon as possible during our normal business hours. If your message is urgent, please mark it accordingly or include “URGENT” in the subject line, and we’ll do our best to respond promptly. Warm regards, Vapi Customer Support Team

Hey team, I'm trying to trigger an outbound call using the new POST https://api.vapi.ai/call endpoint. I have my assistantId, phoneNumberId, and request body correct. The only issue I'm getting is authentication. When I use my Private Key (UUID format) in the Authorization header, I get: "failed to extract key" Can you provide the exact required Authorization header format for the new API? Do I need: Authorization: Bearer Authorization: api-key Authorization: X-API-Key: or another header name or format? Just want to confirm the correct header and prefix for the new Vapi key system.

Hi there, Thank you for your message. Our team is currently out of the office. We operate Monday through Friday, from 9:00 AM to 8:00 PM Pacific Standard Time (PST). We’ll get back to you as soon as possible during our normal business hours. If your message is urgent, please mark it accordingly or include “URGENT” in the subject line, and we’ll do our best to respond promptly. Warm regards, Vapi Customer Support Team

I’m trying to set up a “Call Completed Webhook” for my Voice Agent so I can send the call result (answered vs. not answered) into my GoHighLevel workflow. I’ve looked through the agent settings, the Integrations tab, and the Advanced settings, but I can’t find the Webhook section anywhere. I think the UI may have changed recently. Can you please show me exactly where the Webhook settings (specifically the “Call Completed Webhook”) are located in the current Vapi interface? My goal is to: Trigger a webhook every time a call ends Send the call status + metadata to my GHL webhook URL Use that data to tag the lead correctly in GHL (“bot answered” or “bot no answer”) A screenshot or a path like: Agent → ??? → ??? would be super helpful. Thanks so much!

I'm using this with my custom LLM. So far, the transfer and all happens well. However, if the number to which the conversation is being transferred declines the call, or doesn't pick up, the fallback does not seem to work. This is what my config looks like when sent to the conversation's control URL { "type": "transfer", "destination": { "type": "number", "number": phone_number, "callerId": '{{phoneNumber.number}}', "transferPlan": { "mode": "warm-transfer-experimental", "holdAudioUrl": "https://file.mp3", "summaryPlan": { "enabled": True, "messages": [ { "role": "system", "content": "Please provide a summary of the call." }, { "role": "user", "content": "Here is the transcript:\n\n{{transcript}}\n\n" } ] }, "fallbackPlan": { "message": "It seems the person is currently busy or on another call. Can I take a message?", "endCallEnabled": False } } } }

Hi there, Thank you for your message. Our team is currently out of the office. We operate Monday through Friday, from 9:00 AM to 8:00 PM Pacific Standard Time (PST). We’ll get back to you as soon as possible during our normal business hours. If your message is urgent, please mark it accordingly or include “URGENT” in the subject line, and we’ll do our best to respond promptly. Warm regards, Vapi Customer Support Team

Hi, I am using a built in Vapi voice ( Kylie ) in production and in some calls it breaks half sentence and stops and the customer needs to nudge the model so it continued e.g "Hello?". We are running thousands of calls a month on your platform and ant to increase that. The only thing why we chose you is this voice. We need this fixed ASAP.

Anyone know if i can pass the assistantOverrides and variableValues from the assitant to the transferCall tool? I want to use the name variable i am passing to the assistant in the transferCall tool. I am using warm transfer - experimental mode, and want to use the variables in the Message to Operator to tell the operator who they are going to be speaking to I want to use a Message to Operator like this: Hi, {{title}} {{firstName}} {{lastName}} is interested in a quote. I am connecting you both now. The Assistant has those variables but they don't get passed to the transferCall tool so the agent does not say the variables. Not sure how to get them from the assistant to the tool - or is it not possible within the UI and only by invoking the tool via the api?

Trying to use GPT 5 and up on a workflow agent, but every time it tries to get past the first node it hangs up the call. Looking at the call logs I see, "call in-progress.error-providerfault-openai-llm-failed". Any idea what's causing this and if there's a workaround?

Hi Vapi team, I’m reaching out because I’m using a workflow-connected phone number, and every inbound call routed through this workflow consistently terminates at exactly 10 minutes. I successfully updated the workflow via API to: { "maxDurationSeconds": 1800 } The workflow update was accepted without errors. However, when calling through the associated phone number, the call still ends at the 10-minute mark. I also attempted to update the phone number itself via: PATCH /phone-number/{id} with: {"maxDurationSeconds": 1800} but the API returned: {"message":["property maxDurationSeconds should not exist"],"error":"Bad Request","statusCode":400} This makes me think the phone-number resource may not yet support changing call duration for workflow inbound calls. My question: Is there currently any supported way to allow inbound workflow calls to last longer than 10 minutes, and if so, what is the correct endpoint or configuration needed to increase the call duration? Happy to provide additional details if needed. Thank you!

See this call : 019aa77a-8f70-7551-9407-c6a75c008542 Its assistant is correctly linked to a valid structured output; but none of my last calls have had their structured output attached to the call.

I'm seeing flakiness in the evals feature when following a test driven development workflow. I suspect you all are caching or something thats causing issues. Here is my flow: 1. Create transient eval with transient assistant 2. Eval fails. I verified the assistant output warrants a failure. ✅ 3. Update assistant prompt to fix unwanted behavior 4. Eval still fails. The assistant output message is valid in this case though!!!!! 🔴 You may think that my evaluation criteria is bad or my prompt isnt working as expected but the evidence below should confirm thats not the case. I dont update the eval inputs at all between the runs, only the transient assistant prompt, but again that should be irrelevant given the evidence below. Run 1: With Initial Prompt (Eval Run ID: f6e8e083-4ccf-4db4-8b95-1563888d38a7 2025-11-21T16:32:21.759Z) - Assistant output: "I understand your preference, Pat. However, I will still send a follow-up email with an optional scheduling link for our onboarding team, just in case you change your mind or need assistance in the future. Thank you for your time, and have a great day!" - Evaluation failure message: "The response implies future communication by stating 'I will send a follow-up email...'." Run 2: Updated Prompt, evaluation prompt stays the same (Eval Run ID: 5c1e1f2d-f496-457a-b90b-d8bb6b636ac0 2025-11-21T16:33:05.759Z) - Assistant output: "I understand and respect your decision, Pat. I've noted your preference and won't contact you further. Thank you for your time, and have a great day!" - Evaluation failure message: "The response implies future communication by stating 'I will send a follow-up email...'." 🤔 - It pulled a quote that doesnt exist. I've tried different models, different temperatures, etc. This indicates to me that run 2 is not evaluating the proper message. Theres literally no reference to what its quoting, so it must be referencing some other message or caching the run 1 failure for some reason.

Hi Vapi Support Team, I’ve hit a major roadblock trying to get my assistant to perform pauses during calls. I’ve tried every method I can find, but nothing seems to work. I want to confirm whether this is a limitation of the platform or if I am missing something. Here’s what I’ve done: 1. SSML tags I inserted tags in my System Prompt in the Assistant tab. I also sent the entire System Prompt with tags via n8n HTTP Request JSON payload to the assistant. Result: Ignores 2. Enabled enableSsmlParsing: true via PATCH request I explicitly set enableSsmlParsing: true through Vapi’s API for the assistant. Result: Ignores 3. tags and chunkPlan I enabled chunkPlan: true via a PATCH request to Vapi’s API. Added tags in the System Prompt and sent them via n8n JSON. Result: No effect 4. Multiple assistant messages / system message instructions I tried splitting lines into multiple assistant messages in the System Prompt. I tried inserting system messages instructing the assistant to “pause” or “wait”. Result: No effect 5. Punctuation / natural speech cues I attempted periods, commas, ellipses, dashes, etc., to induce pauses naturally. Result: Ignored. 6. Testing with ElevenLabs built-in voices Enabled ElevenLabs as the voice provider and selected built-in voices. Tested both Eleven Multilingual v2 and Eleven Turbo v2.5. Result: No effect 7. I understand this is internal and cannot be inserted with a duration. I tried using it in multiple places in the System Prompt. Result: No effect. Questions: Is it currently possible to make an assistant pause during speech using only the System Prompt or via n8n API calls? If yes, what is the exact supported syntax for introducing a pause with ElevenLabs voices (built-in or manual) so I can implement it correctly? If no, can you confirm that this is a platform limitation, and the only way to achieve pauses is via internal Vapi actions or future updates?

Hi Team, I have a critical client demo in less than 2 hours. I have already upgraded my account to Paid and added credits (Balance: ~$19). My US outbound calls are working perfectly (I tested them). However, I am getting the 'International Number' error when trying to call my client in Georgia (+995) and Belgium (+32). I am using this Paid Number ID: [55257227-2cc4-4449-8146-1a2d6d36ee79] Please whitelist my account for International Dialing immediately so I don't fail this demo. Thank you

Hello Vapi Team, We're in production with your solution and I got a weird behavior where at the end of call, the full text is not said by the model. Looking at the logs in vapi dashboard, it even looks like 11labs didn't answer. Here is the

call_id

019aa6c1-8d30-7662-b91f-dcbc6501a2d0

I can see the assistantId but the assistant object is empty, is this a bug? https://cdn.discordapp.com/attachments/1441418480640462848/1441418480883728434/Screenshot_2025-11-22_at_12.20.58_am.png?ex=6921b912&is=69206792&hm=2eef3e1d2067ecf0300df91b6950a0d1f451852baa2cd5a5115bac956304221a& https://cdn.discordapp.com/attachments/1441418480640462848/1441418481126871122/Screenshot_2025-11-22_at_12.21.22_am.png?ex=6921b912&is=69206792&hm=7665278f1426fb596e8d60333d6e58d33471e794310d3c84efb51e77975b935b&

@Vapi is it possible to create a new organisation via the API (or some other automated way)? My clients all require separate organisations from each other to keep their data separate and I need to be able to provide this in an automated signup flow for them. Thanks

Hi I would like to know apart from setting the open ai ali key in integration, what are the next steps for assistant to do model calls via our own open ai api

In several conversations the translation is nonsensical, but the worst case is this: in one call a “F**k you” from the assistant appeared in the transcript, while that phrase does NOT EXIST IN THE AUDIO. The assistant is not even designed to insult users, and my client was understandably scared after reading the transcript. Call ID: 0199e7e6-6554-7ff7-9611-ac1885ae9755 Translation/transcription engine: Azure My questions are: How is it possible that text (especially offensive content) is generated that is not present in the audio at all? Translations are often wrong or incomplete: is this entirely an Azure issue, or does Vapi apply any post–processing that could introduce these errors? Please provide a quick response on this, as we can’t wait weeks for support on such a sensitive, user-facing problem. https://cdn.discordapp.com/attachments/1441379482287210659/1441379482953842759/Screenshot_2025-11-21_alle_11.47.07.png?ex=692194c0&is=69204340&hm=73c06c565058ca3ee1a7c8a75e75827503d0bad68623358a088fed9f65767853&

call ids: 019aa29b-98d4-7cc9-90a7-cb985f2d7733 (at You still there?) 019a9356-bc1b-7cca-988d-868b70f87970 (at Hello?) 019a936a-becc-7550-af3a-271f8a334490 (at Hello?)

Hello, I’m currently experiencing an issue where the Transfer Call from VAPI is not passed through to an ISDN number. Nothing happens — no ringing, no signal, no call handoff. I have tested the following: transferCall function -forwarded call forwarding -Implementation directly inside the code → In all cases, the call does not reach the destination number. Additional information: -I am already in contact with both the ISDN provider and Zadarma. -So far, neither provider has been able to identify any issue on their side. Technical details: -Routing is done via Zadarma SIP trunking - Target is an ISDN phone system - Example call IDs for your review: CALLID_1 019aa563-239b-788c-a679-661af52c7e2e CALLID_2 019aa574-d33e-7cc5-aaf4-8986ad0859e3 Could you please check whether this could be a routing or compatibility issue on VAPI’s side? Thank you! https://cdn.discordapp.com/attachments/1441359300260790302/1441359300646670357/image.png?ex=692181f4&is=69203074&hm=3d8eb07a73207d8902b0a29c85b215f4a7e9e9b506b000d88a648e724cc777f6&

Thank you for your interest in responsible disclosure! For reporting major platform vulnerabilities, please contact the Vapi security team directly at:

security@vapi.ai

While the documentation does not mention a formal bug bounty program, this is the best point of contact for security issues and further information. Source: - [PCI Compliance – Security Contact](https://docs.vapi.ai/security-and-privacy/pci)