An operating system
Hears your conversations, turns them into work and runs agents — privately, with no tracking and no single database of your data.
Free forever · no card
### Topic
A walkthrough of the AI assistant for video calls and a discussion of where it can be applied.
### Key moments
- Product demo: User showed a tool that joins the call, transcribes it and gives AI hints in real time. Highlight — a stealth mode the other side never sees.
- Capabilities: Translation across languages, deep web research and model choice (Claude and others) for generating answers.
- Use cases: Interviews, complex negotiations (e.g. legal questions about an LLC) and protecting your own pitch.
- Access: Commercial, aimed at the international market; User offered Speaker 0 a long trial in exchange for feedback.
- Technical: At the end they briefly discussed a four-agent system and working with environment variables.
### Outcome
The call ended on a friendly note; User will grant Speaker 0 test access.
Transcript (171 messages)
User
Over any call — captures system audio
Three steps
From call to a confident answer
Launch before the call
A transparent panel over any app. Nobody on the other side can see it.
Hears and transcribes
Recognizes speech, separates speakers and translates lines when needed.
Suggests the answer
AI offers wording and talking points tuned to your context — resume, role, rules.
Privacy
Your conversations stay yours
Transcription and diarization run locally on your device. Only what is needed for the answer leaves to the cloud.
- On-device local transcription
- Panel is invisible in screen recording and to the other side
- Data is never sold or used for training
Pricing
Start free, pay as you grow
Full feature set on the free plan. PRO — when you need long calls and priority.
Frequently asked questions
By default — no: the panel is visible only to you and stays out of screen shares and call recordings. But it’s your call. When you share your screen, you decide — keep the overlay hidden, or show it on purpose if you’d rather run the conversation openly.
Any of them. SpotMax runs on top of system audio instead of integrating into a specific platform, so it doesn’t care where the conversation happens: Zoom, Google Meet, Microsoft Teams, Telegram, Discord, Slack, Webex, a regular phone call, or even audio in your browser. If you can hear it on your device, SpotMax can process it.
Recognition and translation work for all major languages — English, Russian, Chinese, Spanish, Arabic and more. Translation runs both ways live during the call, and you can get hints in whatever language suits you, even if the other side speaks another one.
Yes — there’s a forever-free plan, no card. It runs local models right on your device: private, with no cloud cost. The free limits are small, so they won’t go far for regular use. Paid plans remove the limits and add cloud models, long calls and priority processing — and you can try a paid plan free for 7 days. Our goal is to keep improving local models so the free tier can do more over time.
With you. Recognition and diarization run on your device, and your conversations aren’t pooled into a shared database — there’s nothing for us to collect, sell or hand over. A single database of every user is the main backdoor, and we simply don’t have one. Cloud processing kicks in only by your explicit choice.
Part of the work — transcription and diarization — runs locally and doesn’t need a constant cloud connection. Translation and some modes still use the network for now. A fully local mode, including translation, is our goal — and the reason we keep developing local models.