Which Gemini model does this skill use?

It uses the stable gemini-embedding-001 model, which supports Matryoshka Representation Learning for dimensions between 768 and 3072.

Is batch processing supported?

Yes, it supports batching, but includes critical warnings about a known ordering bug in large batches and provides a safer chunking strategy to prevent data corruption.

Do I need to normalize the vectors manually?

Yes, if you use any dimensionality other than the default 3072, you must normalize the embeddings for accurate similarity results. This skill provides the helper functions to do so.

Can I use this with Cloudflare Workers?

Absolutely. This skill includes specific patterns for using the REST API within Cloudflare Workers environments where the full Node.js SDK might not be suitable.

What are the rate limits for Gemini embeddings?

The free tier allows 100 requests per minute (RPM). Paid tiers scale significantly higher, up to 10,000 RPM, depending on your billing account and usage history.

Google Gemini Embeddings

Name: Google Gemini Embeddings
Author: evolv3ai

byevolv3ai

0•

데이터 과학 및 ML

Implements production-ready RAG systems and semantic search using Google's Gemini embedding models with Cloudflare integration.

This skill provides a comprehensive guide and implementation pattern for utilizing the gemini-embedding-001 model to generate high-quality text embeddings. It covers everything from basic SDK usage and REST API patterns to advanced batch processing and RAG integration with Cloudflare Vectorize. The skill specifically addresses critical production concerns like Matryoshka Representation Learning, vector normalization requirements for custom dimensions, and documented workarounds for known API batching limitations, ensuring robust AI application development.

주요 기능

010 GitHub stars

02Advanced error handling for rate limits, batch ordering bugs, and memory constraints

03Automated normalization logic for high-performance vector search accuracy

04Support for gemini-embedding-001 with flexible dimensionality from 768 to 3072

05Production-ready Cloudflare Vectorize and Workers integration patterns

06Optimized task types for retrieval, similarity, clustering, and classification

사용 사례

01Building Retrieval-Augmented Generation (RAG) pipelines for AI agents

02Developing high-performance text classification and duplicate detection systems

03Implementing semantic search and document clustering in web applications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add evolv3ai/claude-skills-archive google-gemini-embeddings

For use in Claude.ai and ChatGPT

Download Skill