Vulcan: AI Hub

Vulcan AI Hub brings local & cloud AI, knowledge, tools, automation & device capabilities together in one Android environment.

CHOOSE YOUR AI
- Run supported local models on your device:
- Gemma 4 E2B & E4B
- Phi-4-mini
- Qwen2.5 1.5B
- Qwen3 0.6B
- OLMo-2 1B
- Connect cloud AI through Google Gemini or any OpenAI-compatible API (bring your own key & base URL, e.g. OpenRouter).
- Bring your own API keys, protected with Android Keystore. Local models support CPU/GPU processing with automatic fallback, model configuration & supported features such as thinking mode & speculative decoding.

LEARN
- Matrix Learning lets supported local models work with information you provide.
- Paste text, upload PDF or text files, or fetch supported web content such as Wikipedia pages. In Matrix Mode, the AI answers using the selected material as its source of truth.
- Use Learning to summarize & add that knowledge to the Archive, allowing your local AI to use it beyond Matrix Mode.

REMEMBER
- The Archive Network gives local AI persistent, searchable knowledge.
- Organize records into drawers, create custom drawers, & search using keyword & semantic retrieval. Knowledge learned through Matrix Learning can become part of the AI's broader knowledge base.
- Observations can extract useful information, knowledge topics & AI-generated insights from conversations. These contribute to Vulcan's familiarity with the user & can be enabled or disabled.

ACT
- With Smart Reflexes, supported local AI can interact with your Android device through natural language.
- Control media, flashlight & volume; make calls & compose SMS; create alarms, timers & calendar events; navigate with Google Maps; launch apps; access Android settings; & control Wi-Fi, Bluetooth & Do Not Disturb.
- Reflexes understand contextual follow-ups such as "turn it off" based on the previous action.
- API Uplink allows AI to interact with configured external REST APIs using supported authentication methods & variables.

AUTOMATE
- Scheduled Prompts turn AI into a scheduled assistant.
- Create one-time or recurring prompts—daily, weekly, monthly or yearly—& receive AI-generated responses as Android notifications. Pause or resume schedules whenever you want.

INTERACT
- Chat with streaming responses, image & video attachments, voice input & text-to-speech. See generation stages & performance information, regenerate responses, stop generation & provide feedback.
- Generate images through supported cloud providers, with generated images stored locally. Web image search runs on DuckDuckGo with citations shown inline.

KNOW YOUR DEVICE
- The Diagnostics Deck provides an interactive view of the device running Vulcan.
- Monitor CPU, GPU, RAM, battery, storage, network & thermal state, alongside extensive sensor data. The Deck also provides camera, NFC, Bluetooth, security & permission information.
- A persistent notification can display live CPU, RAM, GPU & device temperature readings while Vulcan is running.

CONTROL YOUR AI
- Directives let you define persistent rules for how your AI should behave. Create, edit & manage them directly through chat.
- Vulcan also provides configurable cloud profiles, model settings, safety controls & session management.

PRIVACY & SECURITY
Vulcan is designed with privacy in mind.
- No analytics
- No crash reporting
- No tracking
- API keys protected with Android Keystore
- Local AI processing stays on your device
- Safety guardrails across generation paths
- Android backup disabled for application data

Local & cloud AI are intentionally different: local models access Vulcan's on-device Archive, while cloud AI keeps a separate cloud memory drawer that never reads the local one. Both can control your device, and SMS always opens as a draft.

Vulcan AI Hub is more than a place to chat with AI.
Give your AI knowledge. Let it remember. Give it tools. Let it act. Schedule it. Connect it to your device.
Give your AI an environment to operate in.