Lario is a real-time virtual microphone for Mac. A native low-latency audio engine sits between your mic and any meeting app, smoothing hesitation and stabilizing your voice while you speak.
No credit card · Apple Silicon + Intel · Works in Zoom, Meet, Teams, Slack, Discord
<20 ms
Live mode adds under 20 ms before your voice reaches the call.
Native
A native Swift engine processes audio on-device, in real time.
2 min
Install, pick Lario Mic in your meeting app, you're live.
Live signal · Engine running
Interview+ preset · engaged
Voice profile
Calm · clear · stable
Output device
Lario Mic
Voice shaping
Adaptive gain · de-esser · envelope
What the engine does
Plays nicely with
System-wide audio
Lario installs as a standard system audio device. Pick "Lario Mic" in your meeting app of choice. No plugin, no extension, no rewiring.
Zoom
Interviews & calls
Google Meet
Client meetings
Teams
Internal syncs
Slack
Huddles & calls
Discord
Streams & comms
Realtime audio engine
Built in Swift on Core Audio, the Lario engine runs the full DSP chain on-device. Pick a preset, dial in latency, and let it shape your voice while you speak.
Five tuned profiles for the moments your voice has to land.
Interview+
Maximum clarity and stability. The default for high-stakes calls.
Interview
Balanced enhancement for everyday meetings and panels.
Confidence
Warm, steady, assertive. Holds your voice forward in the mix.
Clarity
Articulation boost with a subtle presence lift.
Natural
Minimal processing. True to your voice, just cleaner.
Pick the trade-off that fits the call.
Live mode
<20 ms
Sub-20 ms added latency. Tightest possible loop for live conversation.
Studio mode
~85 ms
Full DSP chain including stutter smoothing for prepared speaking.
Platform
macOS 13+ · Apple Silicon and Intel
Stack
Native Swift on Core Audio
Latency range
32 / 48 / 64 / 80 / 96 / 120 ms
System device
Lario Mic (virtual input)
Toggle every stage. Every dial is exposed. Nothing is hidden behind a single slider.
Adaptive gain
Voiced-only loudness control that holds your level near a steady target without pumping silence.
De-esser
Tames sibilance above 5.5 kHz so plosives and sharp consonants stay clean.
Envelope smoother
Lifts mid-sentence volume drops up to +5 dB so you stay heard from start to finish.
Interruption shield
Detects when the other speaker overlaps you and tracks fairness across the call.
Built into the call
Real-time on the call. Insightful after it. Built around a virtual mic that any meeting app already understands.
A native Swift engine processes every chunk on-device. Live mode adds under 20 ms. Fast enough that the call feels exactly as immediate as it always did.
One install, one virtual input device. Pick "Lario Mic" in Zoom, Meet, Teams, Slack, or Discord. No browser extension, no plugin, no special integration.
Interview+, Interview, Confidence, Clarity, Natural. Switch presets in a click. Every dial is still exposed if you want to fine-tune the chain yourself.
After each call, Lario scores pace, filler words, confidence, and stability. And writes a short coach note about how you sounded.
Plug in OpenAI, Anthropic, Google, or OpenRouter with your own key. The desktop app talks to the provider directly. Your prompt never goes through our servers.
Lario was built for stutter smoothing, hesitation reduction, and clearer delivery for anyone whose live voice doesn't always match the idea behind it.
How it works
Three steps. No new tools. No workflow change during the call.
Grant microphone access, install the Lario Mic virtual device, and pick a preset that fits the call.
Open Zoom, Meet, Teams, Slack, or Discord. Set Lario Mic as your input device. Done.
The engine runs in real time. After the call, your post-session insights are waiting in the dashboard.
Post-session AI
When the call ends, Lario turns the session into a structured read on how you sounded. Pace, fillers, stability, confidence. And a short coach note about the patterns it found.
Speaking pace
138 wpm
Inside the ideal 110–160 range
Filler words / min
1.4
Down 32% vs. last 5 sessions
Confidence score
82
Anchored to this session's audio
Stability index
0.91
Voiced loudness held steady
Coach note
“Pace stayed in your ideal range and filler density dropped meaningfully in the second half. The closing answer was the strongest delivery. Keep that frame for next round.”
Lario tracks when the other speaker overlaps you and surfaces a short narrative about how the conversation actually divided across both sides.
Run analysis with your own API key. The desktop app talks to the provider directly. Your key never touches our servers.
Pricing
Run three full sessions on the free plan. Upgrade when Lario becomes part of how you show up.
Annual billing keeps Pro at 50% off the monthly rate.
Free
Run real calls with the engine and the virtual mic before you commit.
Pro
For people who use Lario every week and want every session, every preset, and every insight.
Billed $72/year. Equivalent to $6/month.
FAQ
Direct answers about how Lario actually works, what's free, and what runs where.
Lario is a native macOS app. It runs on Apple Silicon (M-series) and Intel Macs on macOS 13 or newer. A Windows build is on the roadmap and not yet available.
Yes. Lario installs a system audio device named "Lario Mic". Pick it as your input in Zoom, Google Meet, Microsoft Teams, Slack, Discord, or anything else that takes a standard microphone. No plugin, no browser extension, no special integration.
Live mode adds under 20 ms before your voice reaches the call. Studio mode runs the full DSP chain including stutter smoothing and lands around 85 ms. Better suited to prepared speaking than live banter. You pick the trade-off, the engine respects it.
Free includes three full sessions, the real-time stabilization engine, the Lario Mic virtual device, and bring-your-own-key AI insights so you can fully evaluate the workflow. Pro removes the session cap and unlocks the full DSP chain, every preset, full post-session insights, and priority access to new models.
Real-time voice processing is fully native and on-device. No AI provider involved. Post-session insights are optional and use the AI provider you pick (OpenAI, Anthropic, Google, or OpenRouter) with your own API key. The desktop app calls the provider directly, so your key and prompts never go through Lario servers.
The real-time engine processes audio locally on your Mac. Sessions and analysis live in the desktop app's local store. Our backend only tracks lightweight session metadata. Start time, end time, duration, and plan tier. For usage limits and billing. We don't store your audio, we don't sell content, and we don't train models on your calls.
Yes. Cancel from the app or by emailing us. Your access continues until the end of the current billing period. There are no cancellation fees.
Need something more specific? hello@larioai.com
Install Lario, choose Lario Mic in your meeting app, and let the engine handle the rest. Three sessions are on the house.
Setup
Install, pick Lario Mic, start the call.
Engine
Native Swift on Core Audio. Live mode under 20 ms.
