In this session we will learn together the easiest way to get started with AI models, directly in your browser thanks to the new built-in AI in google chrome ( https://developer.chrome.com/docs/ai/built-in ).
We will learn, through live examples and demos how to use the writer, rewriter, translatio...
In this session we will learn together the easiest way to get started with AI models, directly in your browser thanks to the new built-in AI in google chrome ( https://developer.chrome.com/docs/ai/built-in ).
We will learn, through live examples and demos how to use the writer, rewriter, translation, language detection, summarization proof reading and prompt APIs!
Join Google Developer Expert in Web Technologies Victor Sanchez and discover how smart your browser has just become!
DevFest is GDG Surrey’s annual flagship event, bringing together developers, researchers, students, and tech leaders to learn, share, and connect.
This year, we’re exploring the future of AI as a reasoning partner systems that don’t just predict but collaborate, explain, and align with human context.
Whether you’re a coder, designer, researcher, policymaker, or just AI-curious, DevFest 2025 is your chance to discover ideas, build connections, and shape the dialogue on the next era of AI.
Build with AI. Think with AI.
Join us at GDG Surrey DevFest 2025 Virtual Series to imagine what’s next-together.
Size: 3.86 MB
Language: en
Added: Oct 10, 2025
Slides: 42 pages
Slide Content
GDG Surrey
Víctor Sánchez Belmar
GDE in Web Technologies
AI is now
built-IN your
browser!
What are we covering today?
●AI local models… yes! But… why?
●Chrome built-in AI
●7 APIs!
●A lot of hands-on demos!
GDG Surrey
AI everywhere?
maybe… but not for
everyone!
Running your own local
models was already
possible but it was not
easy, until now!
The issues…
●Explosion of local LLMs ‘systems’
●Complex setups (models, weights, quantization)
●HW requirements (drivers, compatibility, etc.)
●Security (trusted sources, sandboxing, data access)
The next big news in
AI will be really
small!
Small local,in-device
models are the next
BIG thing…
GDG Surrey
Browsers are the great equalizer
●Easy access
●Extremely large user base
●Ideal for both consumer devices and integrated
systems.
●This has worked before! (tensorflowJS)
Why would you run your models locally?
Privacy Faster UX Offline
Let chrome
do the heavy
lifting…
Let chrome pick the right expert for the task…
?????? ?????? ?????? ✍ ?????? ✅
Language
Detector
Translator Writer Rewriter ProofreaderSummarizer
Or not…
??????
Prompt
API
Current API statuses
https://developer.chrome.com/docs/ai/built-in-apis
Requirements
⚙ System & Hardware
●Chrome 138 + required (some APIs start from 137 for
Origin Trials)
●Supported OS: Windows 10/11, macOS 13 (Ventura)+,
Linux, ChromeOS (Chromebook Plus)
●≥ 22 GB free disk space, ≥ 4 GB VRAM, unmetered
network recommended
Model download
*translator options will require source/target language too.
What can these
APIs do?
GDG Surrey
Language Detector
-Detect a language without external models.
-Ideal to identify and download the right
translator or correctly tag users posts.
Language Detector
https://developer.chrome.com/docs/ai/language-detection
Translator
-Incredibly powerful translator models for quick
offline use.
-One model per language pair
-Works best in combination with language
detector
-1 shot translation or streaming available
-Implement correction in user input for
grammar, punctuation and spelling.
-It is able to label each one of the changes
-Provides explanations for the mistakes
-Grammarly as an expert model!
ProofReader
ProofReader
-Results formatted as:
Language Model
-Everything gemini nano can do (including
multimodal input).
-Data extraction, ai-search, customization,
explanations, interaction, generation, etc…
Language Model
Language Model
Language Model
Multimodal
GDG Surrey
Let’s have some fun!
DEMO TIME
Ask me anything!
…or later at @VictorSanchez
Q/A
GDG Surrey
Víctor Sánchez Belmar
GDE in Web Technologies
Thank you!