Added llms.txt to Astro Website

Added llms.txt to an Astro site. Since no plugin was found, I asked AI (Cursor/Claude 3.7 Sonnet) to implement it. Introducing that code.

Official Documentation & Resource Links for Major AI Models (Anthropic, Google Gemini, OpenAI)

With new AI models constantly emerging, keeping up with official information can be challenging. This article compiles links to official documentation, pricing pages, and release announcements for major models from Anthropic (Claude), Google (Gemini), and OpenAI (GPT). Use it as a cheat sheet to speed up your development workflow.

Trial and Error Using Japanese with Amazon Titan Foundation Model on Amazon Bedrock

A work log of three snags with Japanese on Amazon Titan / Amazon Bedrock, from model access to console garbling.

Prototyping an Image-Text Similarity Calculator to Measure Image-Text Similarity with Cohere Embed 4

How similar is an image of a requirements document to the text 'EC site requirements'? With Cohere's latest multimodal embedding model Embed 4, calculating such image-text similarities becomes straightforward. This article provides a detailed explanation of the sample code and usage of a Node.js-based Image-Text Similarity Calculator as a prototyping implementation.

CSV Column Auto-Recognition Patterns - Designing and Choosing Between Heuristic and LLM-Based Approaches

Column names from different marketplaces and ERP systems are never consistent. 'product_code', 'item_code', 'SKU' — countless naming variations for the same field. This article compares two approaches to automatic column recognition: heuristic matching with alias dictionaries and flexible LLM-based inference, with practical e-commerce implementation examples.

Self-hosting the OSS Version of Firecrawl

Self-hosting the Firecrawl OSS v1.0.0 with Docker Compose, and trying crawl and scrape locally.

Local LLM Development by Japanese Companies: A Comprehensive Survey of Domestic AI Models

A comprehensive survey of local LLMs developed by Japanese companies, covering technical specs, industry adoption, and cost structures.

Comparing LLM API Gateways — The Traps in Markup, Billing, Regions, and Training-Data Use

Bundling multiple LLM APIs (OpenAI, Anthropic, Google) behind one endpoint has become a standard component of any generative-AI app. Using Vercel AI Gateway as the reference point, this post compares OpenRouter, Cloudflare, Helicone, Portkey, and LiteLLM, plus Bedrock, Azure, and Vertex AI, along four axes: markup, billing model, region, and training-data use. The shared trap: zero markup does not mean zero cost.

Introduction to llms.txt and OSS Marketing Insights

What is llms.txt? It's a file for LLMs to understand websites. We'll explain the benefits of implementation, how to write it, and the procedure. We'll also introduce hints for OSS marketing.

Machine Learning (ML) Recommended Introductory Books Summary

A curated list of introductory machine learning books recommended by engineers at an AI startup, spanning practical, theoretical, and advanced levels.

Mintlify Automatically Generates llms.txt and llms-full.txt

Mintlify automatically generates llms.txt describing LLM content usage policies for documentation sites and llms-full.txt that compiles the entire site into a single file. Learn how these files, which serve roles similar to robots.txt, function.

How to Install Ollama and Use Local LLMs

Ollama is a tool that enables easy setup and operation of large language models (LLMs) in local environments. It ensures data privacy while providing fast responses. Learn how to install Ollama with Homebrew and run models like Llama 3.2 to harness the power of LLMs locally.

The Complete Guide to OSS-Licensed Local Coding LLMs (2026 Edition)

A complete guide to running OSS-licensed (Apache 2.0 / MIT) coding LLMs locally. Includes the latest model comparisons and practical setup instructions as of March 2026.

Surveying Services That Serve OSS Model APIs in a Japan Region — Comparing Them on Data Residency and Coding-Agent Practicality

Terminal-resident coding agents are now everyday tools, and provider-agnostic tools like OpenCode have spread widely. OpenCode Go is an attractive bundle that wraps OSS models for $10/month, but its regions are limited to the US, EU, and Singapore, with no Japan region. From both data residency and latency angles, this post reconciles inconsistent primary sources to map out the services that can serve OSS model APIs inside Japan.

Reading Notes: 'Practical Introduction to Building Chat Systems with ChatGPT/LangChain' by Shingo Yoshida / ChatGPT/LangChainによるチャットシステム構築［実践］入門

Notes on the ChatGPT/LangChain practical introduction — impressions plus the reference links, OSS, and security resources from the book.

Reading Notes: 'How Generative AI Will Change the World' by Shota Imai

Notes from a Matsuo Lab researcher's book on generative AI, metacognition, and 'extreme actions' as a winning pattern.

Speaking at 'Startup Meetup Kumamoto @ XOSS POINT.'

Speaking at Startup Meetup Kumamoto @ XOSS POINT. about a kindergarten search chatbot on Amazon Bedrock.