Verified vocabulary infrastructure for GPTs and the Assistants API
wisdom
⌄
The Scenario
ChatGPT has 300M+ weekly active users. A custom GPT teaches a “word of the day” with verified content — not hallucinated definitions. The GPT calls Word Orb for a deterministic definition with IPA, etymology, and 47-language translations, delivers a structured 5-phase lesson via Lesson Orb, and runs a comprehension check with Quiz Orb. The Assistants API’s function calling makes each Orb endpoint a tool. Same word, same definition, every time — no variance across sessions or users.
1
Word Orb looks up the word
One API call returns a verified definition, 47 translations, pronunciation audio, and etymology.
Loading…
2
Lesson Orb delivers a structured lesson
A 5-phase lesson with the strategist teaching archetype.
Loading…
3
Quiz Orb assesses comprehension
Interactive assessment aligned to the lesson. Try it yourself.
Loading…
4
The Knowledge Graph connects everything
30,288 connections link words to lessons to assessments.
Loading…
Why this matters for OpenAI
GPT function calling + Orb Platform = custom GPTs with verified vocabulary — deterministic content that doesn’t hallucinate definitions, etymologies, or translations
Assistants API tool integration — register Word Orb, Lesson Orb, and Quiz Orb as three function tools with typed JSON schemas for reliable structured output
MCP server at mcp.thedailylesson.com/mcp — ChatGPT’s MCP support means zero-config discovery of all four Orb products as native tools
162,250 words across 47 languages with 60 tone variations — a GPT vocabulary tutor that serves 300M+ users needs content infrastructure, not prompt engineering