The latest model from Open AI, gpt-realtime is their most advanced model, one we were excited to review and implement into some of our existing services where we have been adding and testing voice functionality.
One of the most exciting features is the MCP server support, allowing the voice agent to connect seamlessly with your services and call these tools based on your prompt. Their Realtime API also allows for integration into third party chat services such as our latest chat bot plugin.
Latency has been the biggest issue with voice agents – as well as interrupting the flow as well as seamless, fluid conversations. After testing this mode and the advances they have made, this have been improved significantly, with customisable options to tweak the results. It is now available for any business looking to elevate or streamline customer experience, increase efficiency, and modernise its brand.
We have two UK clients who serve international markets and fast translation is extremely important to them, both have signed up to work with our AI voice services which we have been integrating for them and will be showcasing later in the year.
What is gpt-realtime?
So what is it? Well gpt-realtime is the latest model of advanced speech-to-speech API from OpenAI designed from the ground up for ultra-low latency interactions. It provides developers with the tools to build AI agents that can listen, think, and speak in a seamless, real-time conversational flow, seamlessly switching between language with incredible speed.
Let’s break down the core components that make this possible:
Hyper-Realistic Voices
The first thing users will notice is the audio quality. The voices generated by gpt-realtime are incredibly natural and expressive. They move beyond simple text-to-speech to convey tone and emotion, making interactions feel far more engaging and human. With a range of voices available, businesses can choose a sound that perfectly matches their brand identity. When you are excited you will hear the tone become more excited and if serious the AI voice model will adapt.
OpenAI has a set of voices you can implement into your models to try, ours currently uses one of the only available UK models, but if you are wanting to fully customise your voice, this isn’t currently possible, however our Elevenlabs trials were really positive in recreating any voice imaginable. That is for another article.
WebRTC for Instantaneous Response
The key to a natural conversation is timing. Any noticeable lag can break the illusion and frustrate the user. To solve this, gpt-realtime leverages WebRTC (Web Real-Time Communication). This technology enables a direct, peer-to-peer connection between the user’s browser and the AI, drastically minimizing latency. The result is an immediate back-and-forth exchange, just like a real phone call.
Conversations with Context: Stateful Session Management
A good conversationalist remembers what you’ve said. OpenAI has built robust session management into the API, allowing the AI to maintain context throughout a user’s interaction. It can recall earlier parts of the conversation, understand follow-up questions, and provide coherent, intelligent responses that build on previous dialogue.
Model Context Protocol (MCP)
To make the AI even more aware, the API supports the Model Context Protocol (MCP). This allows for the integration of additional data streams, such as image inputs. For a business, this opens up future possibilities where a customer could not only describe a problem but also show it to the AI agent, leading to even faster and more accurate resolutions.
Update – after the launch of gpt-realtime, more MCP tools have been added,by OpenAi and other developers including, Outlook, Gmail, Google Drive, Teams, Dropbox, Cloudflare Browser, Pipedream, Papal, Stripe and More.
What does this enable? By linking your services via an MCP server to your agent, you enable it to use prompts and conversations to interact with the customer on a deeper level, browsing the site the user is on, navigating them to a set page, emailing the team if they don’t know the answer to a question and even taking payments.
Why OpenAI Voice Agents offer serious benefits for Business
Integrating this technology is more than just a novelty; it delivers tangible business benefits:
- Elevated Customer Experience: Provide instant, 24/7 support that feels personal and natural, dramatically improving customer satisfaction.
- Increased Efficiency: Automate complex inquiries that previously required human agents, freeing up your team to focus on higher-value tasks.
- Enhanced Accessibility: Offer a hands-free way for customers to interact, making your services more accessible to everyone.
- Modernized Brand Image: Position your company as a forward-thinking leader by adopting the latest in AI-driven communication.
- Language – If you offer an international business, you can now have a multilingual chatbot which can engage in most languages, switching seamlessly and providing you with the information back in your language.
Our application: Forerunner® AI Chatbot with gpt-realtime
Understanding the potential of OpenAI’s voice technology is one thing; deploying it seamlessly on your website is another. That’s why we have already integrated the full power of the gpt-realtime API into our Forerunner® AI Live Chat plugin. Try it out for yourselves by opening our chat bot and then pressing the microphone icon. You will then be connected to our AI Agent who you can speak with about our services.
We are excited as this part of the market develops, knowing other players will come along, there are some already but expect responses from Gemini, Grok etc in the coming months or early 2026.
We have already made this OpenAI voice integration available as an add on to our existing free and pro plugin for Forerunner® AI Live Chat bot Pro, you can enable this and benefit from our existing integrations offering text based AI chat and voice, with Webrtc communication all built in. All you will need os the prompt id from OpenAI and it will work seamlessly on any WordPress website.
Core features of OpenAI’s technology to deliver a practical, real-world solution in our live chat system:
- Instant Voice Chat: By leveraging WebRTC, our Voice Add-on for Forerunner® allows your website visitors to speak directly with your AI assistant and receive immediate, natural-sounding responses.
- Seamless Integration: We’ve handled the complex technical work, allowing you to add this advanced voice capability to your site with a simple, easy-to-configure add-on.
The core Forerunner® AI Live Chat plugin is available for free (currently pending full approval in the WordPress repository), providing best-in-class text-based chat. To unlock the revolutionary voice features discussed here, the Voice Add-on is available as a separate purchase.
This technology offers a clear path to enhancing user engagement and improving support efficiency. To see how you can implement a real-time voice solution on your website, explore the Forerunner® AI Live Chat plugin and its dedicated Voice Add-on.
For us we know that Voice chatbots as well as AI in general is something which can divide businesses and it is important to strike a balance. Some companies moving to 100% AI will find their business model may not work as well because people prefer – in most cases, a personal aproach where their needs feel heard by the business. Our own business interacts with most clients either face to face, remote or phone with our customers preferring to work direct with us and providing them with their preferred contact method. Offering our own AI led live chat, which can offer instant resources and initial meetings for example, has seen our own leads change significantly.
We have integrated AI into our workflows to assist and enable faster initial touch points, however the approach will vary business to business, and our personal approach will drive our own use of these services – if they offer value to our clients and ourselves we see this as a great addition, however not something that would take over every aspect of how we work.
If you would like to discuss AI Automation services for your business, let us do the integration for you and contact us today.
