Millis AI
Roadmap
Feature Requests
Changelog

Changelog

Follow up on the latest improvements and updates.

RSS

new

Smarter Tool Responses with Flexible Language Support

In Millis, users can define
phrases for tools
—these are what the agent says when calling a tool.
Previously, the agent
always used the same phrase
, regardless of the language it was speaking, which could cause mismatched responses. Now, agents can intelligently adjust their responses to match the conversation's language.
  • You can now
    set a Response Mode
    when defining a tool.
  • With
    Response Mode set to "flexible"
    , the agent will
    adapt the phrase
    to match its current speaking language.
  • These phrases serve as
    examples
    , so make sure they match the
    language of your prompt
    for the best accuracy.
This makes multilingual interactions
more natural and seamless

new

February Improvements - Faster Calls, Smarter Multilingual Support, and Improved Stability

🌍 Enhanced Multilingual Support
You can now
define a list of languages
for your agent to focus on, meaning
better performance and accuracy
in multilingual conversations.
⚡ Faster Call Start Time
We've optimized the system so your agent can
deliver the first response even faster
, reducing wait time at the beginning of calls.
📞 Improved Call Stability
Improved stability of
call transferring, voicemail detection, and call termination triggers
, ensuring a smoother and more reliable experience.
These updates make your agents even more responsive and reliable—try them out! 🚀

new

Call Transfer Now Supported on Plivo & Twilio

Millis now supports
call transfer on both Plivo and Twilio
, making it easier to
seamlessly hand off calls
to another number.
More provider support is coming soon—stay tuned!

new

Web SDK Handles Tool Calls & Results

Our Web SDK now
receives tool call triggers and results
, allowing the client to handle them directly.
Use case example:
  • A user asks for information.
  • The agent queries via tool call.
  • The agent receives a
    list of options
    as tool call result.
  • The client can now
    display these options
    in the web app for a more
    intuitive user experience
    .
This makes interactions
smoother and more interactive
—try it out! 🚀

new

🗣️ Deepgram Nova-3 Model & STT Model Selection

🗣️ Support for Deepgram Nova-3
Millis now supports
Nova-3
, Deepgram’s latest
state-of-the-art speech-to-text (STT) model
with
improved accuracy
.
  • Supports English for now
    , with more languages coming soon.
🎛️ STT Model Selection
You can now
choose your preferred STT model
for better control over transcription quality. We add two more models from Deepgram:
  • Deepgram "nova-2-phonecall"
    – Optimized for phone conversations.
  • Deepgram "nova-2-conversationalai"
    – Designed for AI-driven conversations.

new

💬 Chat in Dashboard – Faster Testing & Debugging

Xnapper-2025-03-18-00
Great news! You can now
test your agent directly in the dashboard using chat
.
  • Faster iteration
    – Quickly refine your agent without making calls.
  • Easier debugging
    – See responses instantly and fine-tune prompts on the spot.
Try it out and build smarter agents faster! 🚀

new

Redesigned Agent Page – Better UX, Easier Configuration

Xnapper-2025-03-18-00
Great news! We’ve redesigned the Agent Page to make it even easier to use.
  • Improved layout
    for easier navigation and configuration.
  • Larger text fields
    for better visibility and easier prompt editing.
  • All settings neatly organized
    to make adjustments quicker and more intuitive.
Check out the new design and let us know what you think! 🚀

new

⚡ Support for gpt-4o-realtime-mini – Fast & Affordable

  • Now supporting
    gpt-4o-realtime-mini
    , offering OpenAI's
    fast, advanced voice technology
    at a
    lower cost
    .
  • Simply **select
    gpt-4o-realtime-mini
    ** as your LLM option—no extra setup needed.
  • Seamless integration
    with all existing features and workflows.

new

🧠 DeepSeek V3 & Chat History Control

🧠 DeepSeek V3 Model
  • Added support for
    DeepSeek V3
    , the latest
    state-of-the-art open-source model
    .
  • Delivers
    quality close to OpenAI GPT-4o
    at a
    more affordable price
    .
  • Try it out and see how it performs!
📜 Chat History Control
New options to manage chat history to save cost from LLM token usage:
  • Number of latest chat messages to keep
    – Control how much context the agent retains.
  • Include tool call results only from the last N messages
    – Keep responses relevant without unnecessary history.
These updates give you more flexibility and control—test them out and let us know your feedback! 🚀

new

DTMF Dial Support – Navigate IVRs & Voicemail Systems

Xnapper-2025-03-18-00
🔢 DTMF Dial Support
  • Agents can now
    send DTMF tones
    to interact with
    IVR menus, voicemail systems, and other keypad-based inputs
    .
  • Enables seamless
    call navigation and automated selections
    during calls.
🛠️ How to Use
Simply
instruct the agent in natural language
on how to navigate IVR menus, and it will automatically send the appropriate DTMF tones.
For example:
  • "Press 1 for English, then press 3 for customer support."
  • "Press appropriate numbers to reach out to the representative."
No special formatting or manual DTMF commands needed—the agent will handle it for you!
Load More
→