Tools
Small, free, no-signup tools for people shipping LLM systems. Each one exists because I kept rebuilding it for client work. Everything runs in your browser; nothing is logged.
MCP Server Generator mcp-server-generator.html
Describe your tools in a form, get a working Model Context Protocol server in Python or TypeScript, plus the client config for Claude. Opinionated stubs, typed parameters.
LLM Judge Rubric Builder rubric-builder.html
Concrete, checkable eval criteria with 0 to 2 score anchors for seven task types, then a ready-to-paste judge prompt and JSON schema.
Context Window Budget Calculator context-budget.html
Allocate your window across system prompt, tools, retrieval, and history. See it fill, catch the long-context surcharge cliff, and price every call.
RAG Chunking Visualizer chunk-visualizer.html
Paste a document, pick a strategy and overlap, and see every boundary before your embedding model does. Fixed, sentence-aware, and recursive splitting.
LLM API Cost Calculator llm-cost-calculator.html
Your real workload priced across Claude, GPT, and Gemini: volume, cache hit rate, batch discounts, long-context surcharges. Every price editable.
These pair with the essays: the rubric builder comes from Your LLM Evals Are Lying to You, the budget calculator from Context Engineering Is Not Prompt Engineering with a Better Name, and the MCP generator from MCP vs. RAG Is the Wrong Question. Requests and bug reports: LinkedIn.