Welcome to the VelvetSpeak Beta (Read This First)
Subject: ✅ Welcome to the VelvetSpeak Beta (Read This First)
✅ Welcome to the VelvetSpeak Beta
Thanks for helping test VelvetSpeak. This beta is how we turn it into a tool that works in real conversations.
VelvetSpeak is settings-heavy on purpose.
Finding the perfect conversation tool is harder than it looks, so the fastest path to great results is:
1) Use Briefing (Context Lock) before important conversations
2) Start from a Quick Start Pack, then tweak only ONE setting at a time
Docs (start here): https://velvetspeak.com/documentation/ (bookmark this)
==================================================
✅ 5-minute setup checklist
==================================================
1) Confirm your G1 glasses are connected in the Even Realities app
2) VelvetSpeak > Settings > Calibration:
- Run Voice Lock
- Run Mic Calibration
3) Pick a Quick Start Pack (recommended)
4) Before you talk, do a Briefing (Context Lock):
- Talking To (required)
- Goal + Tone (recommended)
- Save Briefing
- After the conversation: Clear Briefing
==================================================
⚠️ Important: Local Model
==================================================
Please keep Local LLM OFF for now unless you are specifically testing it.
My current stack cannot use GPU acceleration for local inference, so it runs on CPU.
This can feel slow and can lead to delayed or broken-feeling suggestions.
When GPU support becomes viable, I will add proper offline support and smarter routing to save requests.
==================================================
🛠️ Service availability (Early Beta)
==================================================
Because this is an early beta, servers may be restarted or taken offline at any time for upgrades, bug fixes, security patches, or development work.
Most interruptions should be brief, and I will give advance notice for planned maintenance when possible.
If the app suddenly stops responding:
- Assume a deploy or restart first, then try again in a few minutes
- If it still fails, post your device details and what you were doing
==================================================
🧪 Feedback that actually helps
==================================================
When you post feedback or bugs, please include:
- Phone model + Android version
- Mode or Quick Start Pack used + key settings
- Whether you used Briefing (and what you set)
- What happened (what was good, what was off, delays, weird tone)
- Repro steps if it’s a bug
- Screenshots help (remove personal info)
Troubleshooting docs: https://velvetspeak.com/documentation/ (see Troubleshooting)
If the docs don’t fix your issue, I may need a 1-on-1 debug session with you during early beta.
==================================================
⭐ Pro features (Beta testing)
==================================================
Pro features I am testing include:
- Access to additional model providers (example: Gemini, Grok)
- Higher-tier smart GPT models for better suggestions
- Early access to experimental features
- Higher request limits
- Debrief Mode (After-Action Review):
When enough information is available, the Debrief indicator turns green and unlocks a deeper debrief.
It analyzes the conversation and suggests which settings to adjust to improve results based on how you talk and your goals.
Note: Each Debrief uses a higher-tier smart model and costs 10x requests per run.
Pro cannot be purchased yet.
Pro will be granted to users who provide high-quality feedback, bug reports, and testing notes.
==================================================
🎥 Demo video reward (1 year of Pro)
==================================================
If you record a real-world demo video and it’s good AND you allow me to use it on my website, I’ll give you 1 year of Pro.
What “good” means:
- Real scenario (date, social, work), not just a settings tour
- Clear audio and readable HUD moments
- Shows the Briefing you used and the settings you ran
- 2 to 6 minutes is perfect
- No personal or private info on screen
==================================================
Final note
==================================================
Yes, the market is small right now because this beta focuses on G1 users.
The plan is to prove what works here, then port to other smart glasses.
Expect early beta issues (crashes, connection hiccups, edge cases).
Report them and we will harden the system fast.
Thanks again. Let’s build the ultimate conversation tool.