Sulyara · Desktop app for Windows
A discreet overlay that listens to your interviews and meetings, reads your screen when you ask, and suggests what to say, grounded in your résumé and the role you're applying for.
Example interview. Suggestions come from the AI model you choose, so check them against what you've actually done.
Start a session before the call. Sulyara stays quiet until you ask it for something.
Your microphone and the meeting audio are transcribed on your own PC with whisper.cpp, so Sulyara always knows what was just asked.
In Local mode, audio never falls back to a cloud service. The default model, base.en, is about 148 MB and downloads inside the app.
"What should I say?" drafts a short reply from the live conversation, using your résumé, the job you're applying for, and the tone rules you set.
Ctrl+Shift+Enter sends a screenshot with the conversation, so it can help with the code, slide, or document in front of you. Screenshot questions can use a separate vision model.
Recap summarizes the call so far at any point. When you stop listening, Sulyara writes a summary, decisions, and action items, and it can recall what was said in your last few meetings.
Sulyara doesn't include AI credits. Add an API key from the provider you prefer and pay them directly for what you use, or point it at a local Ollama server and skip the key entirely.
Your keys are saved on your PC and used only to call the provider or endpoint you configure. The speech model downloads from Hugging Face on first use.
No subscription. You pay your AI provider directly for usage, or run a local model for free.
$10one-time
$20one-time
Checkout and delivery by Gumroad. Choose the version on the next page. Sulyara is free software under the GPL v3; the source code stays public below.
Sulyara has no server of its own. It talks only to the AI provider you choose.
On Windows 10 version 2004 or later, in a session on your own PC, Sulyara asks Windows to exclude its window from screen capture, so it doesn't appear in screen shares or recordings. Over Remote Desktop or in a virtual machine it stays visible and tells you so.
Only when you ask does it send the question, recent transcript, your résumé and job context (and a screenshot for Smart assist) to your AI provider.
Paste your résumé or import a PDF or DOCX, describe the job you're applying for, and set rules for tone and length.
Use a fast model for text and a separate vision model for screenshots. Each answer shows how long the first word took.
Based on cue, GNU GPL v3 or later. Built with Electron and whisper.cpp.
$10 one-time for the app, or $20 with the Build & Use guide. There's no subscription. Because Sulyara uses your own API key, AI usage is billed by your provider, and with a local Ollama model there's no usage cost at all.
You need either an API key from a supported provider (OpenAI, Anthropic, Gemini, Groq, Cerebras, DeepSeek, MiniMax, Azure AI, or any OpenAI-compatible endpoint) or a local Ollama server. Add it in Settings → API Keys.
The installer isn't code-signed yet, so SmartScreen may show "Windows protected your PC". Choose More info, then Run anyway. You can compare the file against the published SHA-256 checksum first.
Transcription runs on your PC with whisper.cpp. In Local mode, audio never falls back to a cloud service. When you ask for help, the question, recent transcript, your résumé and job context, and a screenshot for Smart assist go to the AI provider you configured.
On Windows 10 version 2004 or later, in a session on your own PC, Sulyara asks Windows to keep its window out of screen capture, so it doesn't show in screen shares or recordings. Over Remote Desktop or in a virtual machine it stays visible and says so.
OpenAI, Anthropic, Gemini, Groq, Cerebras, DeepSeek, MiniMax, Azure AI, a local Ollama server, or any OpenAI-compatible endpoint, using your own API keys.
That depends on the rules of the interview, exam, or meeting. Check them before using an AI assistant, and treat suggestions as prompts for your own answer, not a script.
Get the Windows app for $10, add your own API key, and start a session. Questions and feedback are welcome through the contact form.
v0.2.4 · 98 MB · Windows 10 & 11 x64 · App + guide: $20