# Qendryx AI Products Primary site: https://qendryx.com/ Products: https://qendryx.com/#products Sitemap: https://qendryx.com/sitemap.xml Qendryx builds practical AI systems for local inference and business knowledge workflows. The product order is: 1. Lizard — hardware-aware local AI for Windows. 2. SonicMind Connect — meeting intelligence and workflow-ready reports. ## Lizard Product profile: https://qendryx.com/products/lizard/ Official product site: https://lizard-llm.qendryx.com/ Lizard Chat: https://lizard-llm.qendryx.com/chat.html Remote benchmarking: https://lizard-llm.qendryx.com/remote.html Documentation: https://lizard-llm.qendryx.com/docs/ Download: https://lizard-llm.qendryx.com/download/ Lizard profiles CPU, RAM, GPU and VRAM, recommends a compatible GGUF model, verifies the download, and routes work to a suitable local execution path. Lizard Native is the resident Direct3D 12 GPU path. Caterpillar is a separate engine built around a pre-compiled compute graph, typed DAG and reused activation arena. Lizard supports local chat and an OpenAI-compatible local API. Installed models can run offline. The native runner activation is a one-time USD 5 purchase; optional dashboard credits are separate. Lizard Chat is the included local chat workspace. It keeps saved, searchable conversations, answers questions about the user's own text, Markdown, CSV, log and source files with citations, and supports Model Context Protocol plugins for folders, web pages, SQLite files and Git history. When a question calls for one, the model chooses a chart type — bar, line, area, pie, scatter, or a node-link diagram — and the chart is rendered on the device from a charting library bundled in the application, so charts work with no network connection. There is no chart button or special syntax. If the data would produce a misleading chart, Lizard answers in writing and draws nothing. Remote benchmarking is a Pro capability for measuring a second Windows PC. Both computers open outbound HTTPS/WSS connections to a relay, so neither accepts an inbound connection and no port forwarding is required. The host approves the client by signed-in account email and public-key fingerprint before any run. Model weights are never uploaded or relayed. A completed run returns an evidence artifact containing the model id and SHA-256, tokens per second and total milliseconds, prefill and decode split, peak RAM and VRAM, GPU adapter and driver version, and whether a fallback path was used. ## SonicMind Connect Product profile: https://qendryx.com/products/sonicmind-connect/ Product app: https://sonicmindconnect.qendryx.com/ Showcase: https://qendryx.com/showcase/ SonicMind Connect captures meetings, audio, video, browser audio and uploaded files, then creates searchable transcripts, summaries, highlights, action items, follow-up questions and workflow-ready reports. Core SonicMind pages: - https://qendryx.com/features/ - https://qendryx.com/use-cases/ - https://qendryx.com/templates/ - https://qendryx.com/pricing/ - https://qendryx.com/sonic-live/ - https://qendryx.com/showcase/ - https://qendryx.com/why-sonicmind/ Feature pages: - https://qendryx.com/ai-transcription-software/ - https://qendryx.com/ai-meeting-assistant/ - https://qendryx.com/ai-meeting-notes/ - https://qendryx.com/meeting-transcription-software/ - https://qendryx.com/ai-meeting-summary-tool/ - https://qendryx.com/ai-meeting-report-generator/ - https://qendryx.com/browser-audio-transcription/ - https://qendryx.com/ai-action-item-extractor/ - https://qendryx.com/ai-follow-up-question-generator/ Comparison pages: - https://qendryx.com/otter-alternative/ - https://qendryx.com/fireflies-alternative/ - https://qendryx.com/sonix-alternative/ - https://qendryx.com/fireflies-vs-otter-vs-sonicmind/ - https://qendryx.com/otter-vs-sonicmind/ - https://qendryx.com/fireflies-vs-sonicmind/ - https://qendryx.com/sonix-vs-sonicmind/ Canonical public URLs use HTTPS. The legacy /usecases/ route is intentionally excluded.