What's new

You can now generate images without leaving the conversation. Describe what you need, and a real-time streaming preview appears as the image is being created. Once it's ready, the assistant suggests follow-up prompts so you can refine the style, crop, or try a different direction — all without starting over.

After every assistant response, a row of contextual suggestions appears below the message. These are generated based on what was just discussed and help you continue the conversation with a single tap — no need to think about what to ask next.

When the assistant produces longer content — like a blog draft, an email, or a detailed explanation — it now renders in a dedicated content view with an auto-generated title. This makes it easier to scan, read, and reuse compared to inline chat bubbles.

A new switcher in the navigation lets you jump between Chat and Voice instantly. It also supports keyboard shortcuts, so power users can switch products without touching the mouse.

Every chat session now gets an auto-generated title based on the conversation content. This makes it much easier to find past conversations in the sidebar and the documents page — especially when you have dozens of sessions.

When someone calls your voice agent a second time, the agent now remembers the previous conversation. This means callers don't have to repeat themselves — the agent picks up where things left off, including any context about their request, preferences, or prior agreements.

If an inbound call doesn't connect — for example due to a network issue or a timeout — the system now automatically schedules a callback. This ensures no lead or customer inquiry gets lost, even during peak hours.

The entire chat experience has been rebuilt from the ground up. New conversation starters help you get going quickly, the toolbar gives faster access to common actions, and smooth animations make the whole experience feel more responsive. The sidebar has also been reworked to surface your recent sessions and documents more clearly.

The global dashboard has been redesigned around per-agent summary cards. Each card shows key metrics at a glance — recent conversations, call volume, and activity trends — so you can monitor all your agents from one screen without drilling into each one individually.

Every Monday at 08:00 CET, you'll receive an email summarizing your voice agent's activity from the past week. It includes call counts, average duration, and notable conversations — giving you a quick pulse check without logging in.

On mobile devices, the app now has a bottom navigation bar that adapts to what you're doing. When viewing a voice agent, it shows tabs for conversations, settings, and phone numbers. This replaces the previous hamburger menu and makes navigation much faster on the go.

You can now see exactly how many voice minutes each agent has used. Usage is tracked per agent with configurable limits, so you stay in control of costs and can allocate minutes based on each agent's workload.

A new dedicated page lets you manage all your assigned phone numbers in one place. See which number is connected to which agent, and quickly reassign or configure them as your setup evolves.

Choosing a voice for your agent is no longer a guessing game. You can now preview all 30 available voices directly in the setup wizard — listen to each one, compare them side by side, and pick the one that best fits your brand before going live.

Each agent can now have its own voice provider and voice style. If you run multiple agents — say, one for customer support and one for appointment booking — you can give each a distinct personality and tone.

Every voice call now produces an accurate text transcript, available in the conversation detail view. This makes it easy to review what was said, search across calls, and keep a written record of every interaction.

Voice agents can now send text messages during a call. This is useful when the caller needs a link, a confirmation code, or written details they can refer back to later — for example, a booking confirmation or a directions link.

Describe a topic or paste some notes, and the assistant will generate a polished LinkedIn post ready to publish. The output is formatted specifically for LinkedIn — with the right length, tone, and structure — and includes a one-click copy button so you can paste it straight into LinkedIn.

All generated marketing content — ad copy, social posts, product descriptions — now comes with a dedicated copy button. No more selecting text and hoping you got everything. One click copies the full content, ready to paste wherever you need it.

You can now record voice messages directly in the chat. A full-width waveform display shows your recording in progress, making it easy to dictate longer requests or quickly capture ideas when typing isn't convenient.

A new page shows every conversation across all your agents in one unified list. This is especially useful if you manage multiple voice agents and want to review recent calls without switching between agents one by one.

Voice agents can now forward calls to a human when needed. You can configure multiple forwarding destinations per agent — for example, a main office number during business hours and a mobile number after hours — so callers always reach the right person.