Web scraper and token compressor that converts HTML to clean markdown with 70-80% fewer tokens. Single-page compression and multi-page BFS crawling with auto-fallback fetch modes.
LLM routing proxy that cuts API costs 60-90% by auto-selecting the cheapest capable model across OpenAI, Anthropic, and Google. Provides stats, config, and model comparison tools via MCP.
Add a custom term to the voice vocabulary to fix mis-transcriptions. Provide the spelling and optional pronunciation hint so Whisper recognizes it correctly.
Permanently delete a term from your voice vocabulary to stop incorrect corrections. Use for irrelevant or problematic entries. Requires exact spelling. Idempotent—safe to run even if term doesn't exist.
Search the Open Knowledge Graphs catalog for ontologies, vocabularies, and taxonomies. Filter by category or type to find resources for semantic alignment.
Remove a term from a controlled vocabulary when it is not referenced by any catalog records. If the term is in use, the deletion fails with details on how many records reference it.
Get the architecture cheat-sheet with a freshness verdict. See if changed files make it outdated so you can understand the codebase without re-exploring.