THE SOFTWARE 3.0 AMBIENT INTELLIGENCE PLATFORM

Say what you mean.

Get it done at the speed of speech.

The high-bandwidth biological bridge to your Windows PC. Hold Right Alt, speak naturally, and watch dense, pre-polished text paste directly at your cursor in sub-180ms. Follows the Zero-Backspace Rule.

<180ms Latency
Sub-second biological feedback
Zero Focus Loss
Native Win32 caret preservation
Zero-Backspace Rule
Deterministic filler stripping
Secure Local Audio
In-memory 16kHz PCM buffer
Interactive Software 3.0 Sandbox

Test the physical push-to-talk mechanic.

Hold the button or press and hold Space to experience the exact sub-180ms flow with synthetic audio feedback. Follows the Zero-Backspace Rule.

WinKeyServer IPC: Connected|
Foreground Target Window:
cursor : auth.middleware.ts
VESPER Ready · Hold Alt
// Composer Prompt (Ctrl+I)Caret Focused in Foreground Window
Press and hold the button below to stream spoken intent here...
Zero-Backspace Rule Active
Validated User Ergonomics

Engineered for high-frequency daily workflows.

Not generic speech-to-text. Tested and certified across 5 acute jobs-to-be-done with the Zero-Backspace Rule.

Alex Chen

Senior Full-Stack Engineer

AI Coding in Cursor Composer
"I formulate complex code prompts at 180 WPM without ever touching the mouse or losing my cursor in Composer."
Primary Tools:

Cursor, VS Code, Windows Terminal

Hardware Setup:

Mechanical Keyboard, Windows 11 Dev Box

Target AppCursor Composer
Latency172ms
Focus Preserved100%
Raw Spoken Audio Input
With verbal pauses & missteps

"Refactor this auth middleware to check for bearer tokens in the authorization header and return a four zero one if missing."

VESPER Deterministic Polish & Caret Injection
Target Caret Output (Zero Backspace)
Sub-180ms

Refactor this auth middleware to check for bearer tokens in the Authorization header and return a 401 if missing.

8-Layer System Architecture

Engineered as a native Windows power-tool. Not another browser wrapper.

Discover how VESPER achieves sub-220ms end-to-end voice-to-action execution while guaranteeing zero window focus loss.

End-to-End Latency Waterfall (<220ms Total)
Click steps to inspect mechanics
Native Win32 Hook

Hold-to-Talk: Sub-180ms Trigger

Press and hold Right Alt (or F8 / Fn) for ≥180ms. The non-activating obsidian capsule floats at the top without stealing window focus.

Release the key to transcribe via the high-speed neural engine and paste directly at your exact cursor position across any active Windows application.

Cursor ComposerVS CodeGoogle ChromeSlack DesktopMicrosoft ExcelObsidian VaultWindows Terminal
Port 32123

VESPER Context Protocol MCP Server

VESPER runs an embedded Model Context Protocol (MCP) server. External AI agents in Cursor, Claude Code, Windsurf, and Antigravity can directly query active desktop memory.

http://127.0.0.1:32123/mcp

Exposes search_user_context and get_active_screen_context with zero cloud transmission.

380ms Window

Double-Tap Persistent Mode

Double-tap Right Alt within 380ms to lock the HUD open for long-form dictation, meetings, or multi-step prompt drafting.

Features interactive on-screen controls for Record, Pause / Resume, Stop & Insert, and Discard.

Dual Pipeline

Direct Dictation vs. Smart Copilot

DIRECT MODE (VERBATIM)

Pure verbatim transcription with strict capitalization, punctuation, and code formatting. Zero hallucinations.

SMART COPILOT MODE

Contextual prompt engineering designed specifically for Claude 3.7, ChatGPT, and Cursor Composer workflows.

Strict HITL Gate

Autonomous Screen Copilot & Safety

Multimodal visual inspection of active windows via low-latency native screen capture.

HUMAN-IN-THE-LOOP SAFETY GATE

Actions with external side effects require explicit approval: press Enter to execute or Esc to reject.

DirectML & CUDA

Hardware Acceleration (Jensen's Law)

Co-designed for modern AI PC silicon. Operates on dedicated NVIDIA RTX Tensor Cores, Intel Core Ultra NPUs, or AMD Ryzen AI via DirectML execution providers.

Runs quantized INT8 models under 60 MB RAM with less than 1W background power, keeping laptop thermals whisper-quiet.

KEYBOARD ARCHITECTURE

Global Shortcuts & Hardware Keycap Matrix

Zero-delay low-level Win32 hardware hooks. Works across any active window without stealing caret position.

Live Win32 Virtual Key Tester
Press any key to test native hook
HoldRight Alt≥180ms
Transcribes audio on release and pastes directly at active cursor without losing focus.
Double-tapRight Alt≤380ms
Locks the floating capsule open with manual Record, Pause, Stop, and Discard controls.
CtrlShiftSpace
Summons the full workspace overlay, intent history, memory hub, and settings panel.
Esc
Instantly aborts in-progress recording without pasting, or closes active modals.
123
Selects numbered clarification pills when the AI detects ambiguous instructions.
Transparent Unit Economics

Free on-device compute. High-leverage Pro cloud.

VESPER never holds your local PC hostage. Use unlimited on-device Whisper for zero dollars, or unlock sub-180ms neural cloud speed and VESPER Context Protocol MCP memory.

Free Tier

30m Daily Allowance
$0/daily quota resets at 12:00 AM

Essential ambient voice dictation with 30 minutes daily usage limit for individual developers.

Included Capabilities
  • 30 minutes daily voice dictation
  • Automatic quota reset at 12:00 AM midnight
  • Silero Voice Activity Detection (VAD)
  • Win32 zero-focus caret pasting
  • Sub-5ms on-device vector context search
  • Local SQLite history & session memory
Download Free Installer

VESPER Pro

Instant License
$8/per month

Or $99 one-time Founder Lifetime Pass

Unlimited ambient dictation, high-speed Neural LPU pipeline, and priority cloud acceleration.

Included Capabilities
  • Unlimited daily dictation (no 30m cap)
  • Sub-180ms ultra-low latency Neural LPU pipeline
  • Instant license generation & automated key delivery
  • In-place Selection Transformer (Edit Mode)
  • Multi-utterance context chaining & memory
  • Custom vocabulary & voice expansion snippets
  • Automatic multilingual speech-to-English letters

Team & Enterprise

Enterprise SLA
Custom — tailored SLA & seats

Dedicated infrastructure, custom fine-tuned Whisper models, centralized team seat management, and zero-retention privacy for engineering teams.

Included Capabilities
  • Everything in VESPER Pro
  • Dedicated enterprise Neural LPU & local GPU clusters
  • Centralized team dictionary synchronization
  • SOC2 Type II & HIPAA compliant zero-retention
  • Centralized admin console & license seats
  • 24/7 dedicated support engineer & SLA

Zero Cloud Audio Lock-In Guarantee

Both Free and Pro tiers include complete access to on-device ONNX execution. If you ever cancel Pro, VESPER remains 100% operational on your local machine with zero downtime.

Hardware Verified:RTX CUDADirectMLIntel NPU

DOWNLOAD & ONBOARDING

Zero-friction setup. Running in under two minutes.

Download the pre-compiled native installer, activate your VESPER subscription license, and begin speaking anywhere in Windows.

VESPER for Windows

Official Release

~170 MB · Windows 10 & 11 (x64) · Direct Caret Injection

Verified Release Package
SHA-256 Checksum Verified
Notice: Because this is an unsigned community preview, Windows SmartScreen may show an initial prompt. Click More info > Run anyway.

VESPER authenticates desktop sessions against your developer subscription license. Sign in with Google or verify your email on this portal to generate your license key.

  1. Create your account or sign in using the top navigation bar.
  2. Verify your email address via the 6-digit PIN sent to your inbox.
  3. Copy your assigned license key (format: VESPER-XXXX-XXXX-XXXX).
Open Guidance & License Gate

Product Trajectory

Engineered for privacy. Built for speed.

From today's live Windows MVP to a 100% air-gapped, offline ambient intelligence platform.

PHASE 1: CURRENTProduction v1.0 Active

Windows Native Push-to-Talk & Context Engine

Full native Windows client with sub-180ms neural speech pipeline transcription, zero-focus cursor pasting, and direct verbatim dictation.

Hardware-hooked Right Alt via WinKeyServer.exe
Win32 foreground window and caret preservation
Direct Dictation and Smart Copilot prompt modes
Human-in-the-Loop safety approval system
PHASE 2: UPCOMINGIn Active Build

Air-Gapped On-Device Whisper (ONNX + Transformers)

Zero internet connection required. Run compact Whisper models entirely on local GPU and NPU hardware with INT8 quantized weights.

Local ONNX Runtime engine execution
NVIDIA TensorRT and Intel NPU acceleration
Zero egress network guarantees for classified workflows
Local vector semantic index for clipboard memory
PHASE 3Planned

macOS & Linux Companion Architecture

Cross-platform parity with global accessibility hooks, native Wayland/X11 support, and per-app prompt templates.

macOS Apple Silicon (Metal) native binary
Application-specific prompt presets (Cursor, Slack, Word)
Global multi-monitor capsule positioning
Unified local telemetry and history sync
PHASE 4Architecture Roadmap

Team Dictionary & Enterprise Vocabulary Sync

Collaborative domain dictionaries for engineering organizations, medical terminology, and proprietary codebase symbols.

Real-time synchronized custom vocabulary
Enterprise audit logging and compliance verification
Self-hosted team dictionary relays
Role-based access controls for corporate deployments

PRIORITY PIPELINE

Join the VESPER Pro & Multi-Platform Beta

Be the first to access on-device offline Whisper models, team custom dictionaries, and cross-platform companion releases. All submissions are indexed into an Excel-ready database.

Platform Interest
Knowledge & Governance

Straight answers from the builders.

Architecture details, security guarantees, and execution protocols.