Personal AI for the glasses you already use.
CyanBridge / Android + Model Studio
Connect supported smart glasses to models you control—on your phone, your computer, or infrastructure you explicitly choose.
Connect supported glasses and manage everyday AI workflows.
Run locally, point to a self-hosted server, or explicitly select cloud inference.
Review data controls and inspect the public Android source.
The product, in practice
The glasses are the interface. The phone is the bridge. You decide where the model runs and which capabilities are allowed to join the workflow.
Check device compatibility01
Connect
Pair supported hardware, manage media and image workflows, and move between glasses without handing your personal AI context to one hardware vendor.
Connection status, media controls, voice and image questions
02
Choose compute
Download a curated on-device model, import a compatible package, or connect to an OpenAI-compatible server. Cloud inference remains an explicit choice, never a hidden default.
Local model packages, runtime controls, self-hosted endpoints
03
Extend
Use community and Tasker integrations for accessibility, productivity, translation, and agent supervision. Support status stays visible so experimental paths are not presented as finished features.
Inspectable plugins, explicit permissions, compatibility labels
Privacy, made visible
CyanBridge keeps local and remote paths distinct. You can see what is selected, where it runs, and when a service outside your devices is involved.
Capture only through supported hardware controls and permissions.
Routes media, questions, and selected app data through an inspectable client.
Local models, datasets, and privacy filtering remain under your control.
Used only when you choose a hosted model or explicitly launch a quoted Studio job.
Import selected exports, control the memory vault, and clear local data from the app.
Inspect device integration work and current implementation details instead of relying on a badge.
Hosted inference, when useful
Mobile plans cover hosted model responses in the Android companion. Local and self-hosted use remain available, while Model Studio cloud jobs are separately quoted with a budget reservation and hard cap.
View mobile plans3.5M reference tokens
$1 / mo
For testing the mobile companion and occasional hosted requests.
18.5M reference tokens
$5 / mo
For regular hands-free questions, updates, and mobile workflows.
74M reference tokens
$20 / mo
For frequent use and higher-volume mobile or agent activity.
Ready when you are
Essential answers
Need device-specific help or billing support? Visit the support center for detailed guidance.
Visit supportCyanBridge combines an Android companion for supported smart glasses with Model Studio workflows for personal datasets, local models, optional cloud compute, and supervised coding-agent work.
The Studio workflow can keep raw exports, parsing, privacy filtering, dataset review, model training, and inference on your machine. The Android companion can also use compatible on-device models or a server you control.
Support varies by hardware, firmware, APIs, and permissions. CyanBridge documents stable, experimental, and research paths separately for HeyCyan-compatible devices, audio glasses, Meta Ray-Ban, MemoMind, and other integrations.
Yes. You can import compatible local model packages or connect the companion to an OpenAI-compatible endpoint you operate. Hosted CyanBridge inference is available as an option, not a requirement.
The Android companion and device-integration work are published on GitHub. Model Studio is maintained separately and is not represented as publicly available source.
Monthly plans cover hosted model responses for the mobile companion. Optional Model Studio cloud training and inference jobs are separately quoted, reserved, and capped before launch.