Pricing

CyanBridge mobile AI plans

Monthly plans cover hosted model responses through the CyanBridge mobile companion — $50 CyanBridge glasses are sold separately before tax/shipping via signed quote vs Ray-Ban Meta ~$379+ and Omi ~$89 + $16/mo. Reference tokens are a shared usage unit based on model cost, not a fixed number of messages. Fully local AI via LiteRT (Gemma 4) / llama.cpp (Qwen3.5) on the phone or Ollama/Model Studio on a laptop needs no subscription. Local Model Studio workflows remain local-first, while optional cloud compute is quoted separately.

Cheap

$1/month

For trying the CyanBridge mobile companion and occasional requests.

What's included

  • Hosted model responses through the mobile companion
  • Up to 3,500,000 reference tokens/month
  • Standard support
Subscribe — $1/mo
Recommended

Standard

$5/month

For everyday hands-free prompts, updates, and mobile workflows.

What's included

  • Models selected for the CyanBridge mobile companion
  • Up to 18,500,000 reference tokens/month
  • Priority support
Subscribe — $5/mo

Max

$20/month

For frequent mobile use and higher-volume agent updates.

What's included

  • Highest mobile inference quota
  • Up to 74,000,000 reference tokens/month
  • Priority support
  • Early access to new models
Subscribe — $20/mo

What $50 includes — and what works fully local with no subscription

Hardware

$50 before tax/shipping per pair (black frame, signed quote shows $50 + freight + estimated taxes + delivery window in BRL authoritative at payment). Ray-Ban Meta Gen 2 ~$379, Display ~$799 for anchor.

Software — optional or $0 local

$1/$5/$20 plans for hosted responses, or $0 local: LiteRT (Gemma 4 E2B 2.58GB/E4B 3.65GB) + llama.cpp (Qwen3.5 0.8B 563MB) on phone (GPU/CPU, templates qwen_chat/gemma_it), or Ollama http://localhost:11434 / llama.cpp / Model Studio on a laptop. 9 native plugins and Obsidian sync work locally.

Local requires checking DeviceCapabilityService (RAM ≥4GB for starter, ABI arm64-v8a/x86_64, free storage). MYVU: display/mic plugins only. NPU Gemma 4 E2B coming soon for Snapdragon.

Optional Model Studio cloud compute

Build and fine-tune locally by default. When you choose paid cloud fine-tuning or hosted inference, Studio presents an itemized quote, budget reservation, and hard spending cap before launch.

30-day mobile trial

New users can try lower-cost mobile AI models without a payment method.

Start free trial

Mobile subscriptions are billed monthly in USD through Asaas or Paddle, depending on checkout availability. Prices exclude applicable taxes. Optional Model Studio cloud jobs are quoted separately. Direct hardware orders, when offered, use their own checkout, shipping details, terms and fulfillment flow. See our Terms of Use and Refund Policy.