Trimming Gemini Embeddings from 3072 to 768 Dimensions: A Matryoshka Approach to Cutting Vector DB Cost and Latency
gemini-embedding-001 returns 3072-dimensional vectors, but thanks to Matryoshka representation you can keep only the leading dimensions with almost no quality loss. This is a design for trimming to 768 to cut vector DB storage and latency, including the re-normalization pitfall and coarse-to-fine search code.
Coalescing Gemini API Requests with SSE Fan-out: Collapsing 100 Simultaneous Hits into a Single Call
How I rebuilt the post-push-notification thundering herd on a 50M-download wallpaper app into a Cloudflare Durable Objects coalescer with SSE fan-out, cutting Gemini API costs by 92% with 14 days of production telemetry.
Apple Vision Framework × Gemini API: Hybrid Image Recognition — Cutting Wallpaper App Cloud Inference Costs by 70%
How I built an on-device prefilter with Apple Vision Framework to cut Gemini Vision API calls by more than half in my iOS wallpaper app. Real cost, accuracy, and latency numbers, with the gotchas an indie developer hits along the way.
Putting Gemini 2.5 Flash Thinking Mode to Work: Reading the Cost-Accuracy-Speed Tradeoff
After three months of testing Gemini 2.5 Flash's Thinking Mode on real projects, here's what actually works: which tasks benefit, which tasks waste budget, and how to build a cost-aware switching layer.
Cut Gemini API Costs by 6x with Gemini 2.5 Flash-Lite
Gemini 2.5 Flash-Lite is now stable and generally available. This guide compares pricing against Flash and Pro with real numbers, walks through Python code examples, and explains which tasks are a perfect fit—and which aren't.
Run Gemma 4 Locally: Building Zero-Cost AI Apps
A complete guide to running Gemma 4 locally and building AI apps with zero API costs. Covers setup on MacBook and Raspberry Pi, production architecture design, and leveraging the 256K context window.
How I Cut My Gemini API Bill from ¥52,000 to ¥8,400 a Month — Caching, Model Routing, and the Batch API
A working record of cutting my Gemini API bill from ¥52,000 to ¥8,400 a month. Covers implicit vs. explicit caching, Flash/Pro routing rules, migrating to the Batch API, and a usage_metadata logging setup — with the production code I actually run.
Gemini API Context Caching— Cut Document Processing Costs by 90%
Learn how to use Gemini API's context caching to reduce repetitive document processing costs by up to 90%. Includes Python SDK implementation, caching strategies, and cost calculations.
Gemini Batch Processing API Guide— Process Thousands of Requests at 50% Off
A comprehensive guide to Gemini's Batch Processing API. Learn how to process thousands of requests asynchronously, cut costs by 50%, and build production-grade batch pipelines with Python and TypeScript.