Feature Requests

OpenAI just introduced the new gpt-realtime and Realtime API updates for production voice agents. This is different from the previous preview realtime models and should be adopted by VAPI ASAP. https://openai.com/index/introducing-gpt-realtime/

Hi. I just added API key for OpenRouter which was successful and now it is shown in Connected section in Integrations. However, when go to Assistants->[Assistant]->Model->Model Settings and choose OpenRouter as provider - i cannot select model, Model drop down menu is not working. When switching to Old Layout, drop down menu was not still available for OpenRouter. But I could paste model id manually and it works this way. When switchhing back to New Layout, drop down still not available in Model Settings, but shown in Model block.

please add openai/gpt-chat-latest (GPT-5.5 Instant)

I tried several cloned Minimax voices but without any error in VAPI the voice just remains silent. d2b0702a-a139-47a1-9f0b-52f7c2894a7b User: martin.buchberger@homebay.ch

Request WhatsApp capability for text chat with Vapi in addition to SMS inbound. WhatsApp can be logged for compliance purposes in regulated industries whereas SMS cannot. Enabling WhatsApp would be material for some client types.

I received your email regarding suspicious activity on my account. I want to assure you that all activity on this account is legitimate. My name is Carlos and I am the owner of Business Pros Inc, a commercial loan brokerage based in Atlanta, Georgia. We are in the process of setting up an AI voice assistant to follow up with business loan applicants. Our developer has been configuring and testing the Vapi integration, which may have triggered your fraud detection. I am requesting an immediate manual review of my account so we can continue building our system without interruption. Please let me know what information you need to verify my account and restore full access. Thank you, Carlos Business Pros Inc info@businessprosinc.com

Our engineers are integrating with your APIs and we meet some problems. Could you please help to to check how to solve these problems? Model parameter error or insufficient credit/quota. Did the https://api.vapi.ai/chat interface return nothing? Also, the request timed out. The response from https://api.vapi.ai/chat is too long and contains too much irrelevant content. How can I debug this? The response from https://api.vapi.ai/chat is too slow; it takes 10 seconds to reply. Using Provider: openrouter, Model: mistralai/mistral-large. Business flow: Client → Backend → /chat API request, waiting for /chat API response. Do you have technical support teams that can answer our questions? Looking forward to hearing from you. Thank you!

Our engineers are integrating with your APIs and we meet some problems. Could you please help to to check how to solve these problems? Model parameter error or insufficient credit/quota. Did the https://api.vapi.ai/chat interface return nothing? Also, the request timed out. The response from https://api.vapi.ai/chat is too long and contains too much irrelevant content. How can I debug this? The response from https://api.vapi.ai/chat is too slow; it takes 10 seconds to reply. Using Provider: openrouter, Model: mistralai/mistral-large. Business flow: Client → Backend → /chat API request, waiting for /chat API response. Do you have technical support teams that can answer our questions? Looking forward to hearing from you. Thank you!

Cannot connect a phone number from a Telnyx sub-account. Telnyx is connected to VAPI using an integration API key (not SIP). When attempting to add a phone number from a sub-account to VAPI, a 404 error occurs. This happens because VAPI is connected to the main Telnyx account, not the sub-account. Proposed solution: Similar to Twilio, allow specifying a telnyx_api_key instead of credentialId when adding Telnyx phone numbers. ___ Without this change, Telnyx sub-accounts are unusable in VAPI, despite being a key feature for enterprise setups.

Hi VAPI team, I'd love to see native support for Mistral's Voxtral model family as a provider option across all three pipeline layers: • STT: Voxtral Realtime (sub-200ms latency, 13 languages, open-weights Apache 2.0) • LLM: Voxtral Small 24B (function-calling directly from voice, multilingual) • TTS: Voxtral TTS (4B params, very low latency, 9 languages, zero-shot voice cloning) What makes Voxtral particularly interesting for VAPI users is that it's the first open-source stack that covers all three layers under a single model family Voxtral Realtime already achieves near-offline accuracy at 480ms delay, which fits right within VAPI's <700ms voice-to-voice target. The Apache 2.0 license also makes it attractive for enterprise and regulated industry use cases. Would love to see Mistral API keys supported at minimum, and ideally a self-hosted endpoint option for Voxtral Realtime via vLLM. Thanks very much guys !