← ClaudeAtlas

web-ai-prompt-apilisted

Chrome's built-in Prompt API implementation guide. Use Gemini Nano locally in browser for AI features - session management, multimodal input, structured output, streaming, Chrome Extensions. Reference for all Prompt API development.
aiskillstore/marketplace · ★ 329 · AI & Automation · score 79
Install: claude install-skill aiskillstore/marketplace
# Chrome Prompt API Reference Complete implementation guide for Chrome's built-in Prompt API using Gemini Nano. ## Hardware & OS Requirements **OS:** Windows 10/11, macOS 13+ (Ventura+), Linux, ChromeOS (Platform 16389.0.0+) on Chromebook Plus **Storage:** 22 GB free (model downloaded separately) **GPU:** >4 GB VRAM OR **CPU:** 16 GB RAM + 4 cores **Network:** Unmetered connection for download **Chrome:** 138+ (Extensions in stable, Web in origin trial) **Not supported:** Mobile (Android, iOS), non-Chromebook Plus ChromeOS Check model size: `chrome://on-device-internals` Model removed if storage <10 GB after download. ## Language Support From Chrome 140: English, Spanish, Japanese (input/output) ## Availability Check ```javascript const availability = await LanguageModel.availability(); // Returns: "unavailable" | "downloadable" | "downloading" | "available" ``` **CRITICAL:** Always pass same options to `availability()` as you use in `create()`. Some models don't support certain modalities/languages. ## Model Parameters ```javascript await LanguageModel.params(); // { defaultTopK: 3, maxTopK: 128, defaultTemperature: 1, maxTemperature: 2 } ``` ## Session Creation ### Basic Session ```javascript const session = await LanguageModel.create(); ``` ### With Custom Parameters ```javascript const params = await LanguageModel.params(); const session = await LanguageModel.create({ temperature: Math.min(params.defaultTemperature * 1.2, 2.0), topK: params.defaultTop