# Add Gemini Nano

> Add on-device LLM via Gemini Nano (Pixel 8+, Galaxy S24+) using MediaPipe LLM Inference or ML Kit. Use when the user says 'gemini nano', 'on-device ai android', 'on-device llm android', 'no-cloud llm'.

- **Type:** Skill
- **Install:** `agentstack add skill-khadinakbarlabs-expo-mobile-app-builder-add-gemini-nano`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [khadinakbarlabs](https://agentstack.voostack.com/s/khadinakbarlabs)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [khadinakbarlabs](https://github.com/khadinakbarlabs)
- **Source:** https://github.com/khadinakbarlabs/expo-mobile-app-builder/tree/main/skills/add-gemini-nano
- **Website:** https://khadinakbar.com

## Install

```sh
agentstack add skill-khadinakbarlabs-expo-mobile-app-builder-add-gemini-nano
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Add Gemini Nano (On-Device LLM)

Android's equivalent of Apple Foundation Models. Free, offline, private.

## Hardware required
- Pixel 8 Pro / 8a / 9 / 9 Pro
- Samsung Galaxy S24 / S25 series
- Some OnePlus 12+
- ~4GB+ RAM for model

## Options

### Option A: MediaPipe LLM Inference (recommended)
```bash
# Add to expo prebuild via config plugin
pnpm add @1mt/expo-on-device-ai
```

```tsx
import { generateText } from '@1mt/expo-on-device-ai';
const result = await generateText({ prompt: 'Summarize: ...' });
```

### Option B: AI Edge SDK (Google native)
Requires native Kotlin module — drop down via expo-modules-core.

### Option C: ML Kit GenAI (newer, 2025+)
For specific tasks: summarization, image description, etc.

## When to use vs cloud LLM
- Privacy-sensitive (no data leaves device)
- Offline functionality required
- Cost reduction (no API bill)
- Avoid Google Play AI policy disclosure (no third-party transfer)

## Cross-platform pattern
- iOS: Foundation Models (skill: `add-foundation-models`)
- Android: Gemini Nano
- Wrapper: `@1mt/expo-on-device-ai` covers both with same API

## Limitations
- Smaller context window than cloud (~1k-4k tokens)
- Generation speed: 10-30 tok/sec
- Only ~50% of Android devices support
- Quality below GPT-4 — good for summarization, classification, simple chat

## Fallback for unsupported devices
```tsx
if (await isGeminiNanoAvailable()) {
  return await generateText({ prompt });
}
return await callCloudLLM({ prompt });  // with 5.1.2(i) consent
```

## Pair with
- `add-openai-streaming-android` for cloud fallback
- `play-ai-disclosure` for compliance (still needed even for on-device if it's AI-generated content)

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [khadinakbarlabs](https://github.com/khadinakbarlabs)
- **Source:** [khadinakbarlabs/expo-mobile-app-builder](https://github.com/khadinakbarlabs/expo-mobile-app-builder)
- **License:** MIT
- **Homepage:** https://khadinakbar.com

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-khadinakbarlabs-expo-mobile-app-builder-add-gemini-nano
- Seller: https://agentstack.voostack.com/s/khadinakbarlabs
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.