From first phonics through doctoral research — built on Koydo's own teaching and the best openly-licensed academic sources in the world.
Koydo's 814 in-house curriculum modules — medicine, law, MBA, computer science, engineering, humanities, K-12 — combined with 111+ openly-licensed academic sources (MIT OpenCourseWare, Open Yale, OpenStax, Khan Academy, PhET, NASA, Wikimedia, and more).
Meridian is the knowledge base — what there is to learn. Koydo Cortex is how it reaches you: personal learning paths, adaptive practice, an AI tutor that notices where you're stuck. Every Cortex-driven recommendation pulls from Meridian.
Meridian is built on four orthogonal axes. Every concept is placed at the intersection.
Meridian is organized by learner stage, not age. A career-changer learning Python is at the same stage as an undergraduate taking their first CS course. Stages line up with the international classification used by world university rankings and partner institutions.
Every concept belongs to one of these fields. We use the same classification as UNESCO and world university rankings, so partner institutions and accredited programs line up directly with Meridian data.
The world's strongest universities share a common general-education core. Meridian requires the same six pillars of every learner, regardless of stage or field.
Hard problems — climate, AI ethics, democracy, health — that can't be solved inside one discipline. We tag concepts across fields so learners can follow a challenge all the way through.
Meridian is promoted by a DB-side function that enforces every gate. Content never reaches the public surface without meeting both review and corroboration requirements.
Edge-cached read-through snapshots served by Cloudflare. Every response carries an ETag and aggressive Cache-Control. Designed for 10M DAU.
| Endpoint | Returns |
|---|---|
| GET /v1/concept/{slug} | Full concept bundle with prerequisites, unlocks, related, content_summary. |
| GET /v1/domain/{code} | Domain overview (D1–D14) with top concepts. |
| GET /v1/age-band/{code} | Age-band rollup (early-years through senior) grouped by domain. |
| GET /v1/search?q=… | Semantic + full-text search via pgvector HNSW. Sub-second on 10K+ concepts. |
| GET /v1/index | Global slug → canonical_uri map for client-side routing. |
| GET /v1/contracts/{ver}/{kind} | JSON Schema for each content kind — clients validate before POST /ingest. |
| GET /v1/health | Liveness probe + stats (concept count, snapshot freshness, active sources). |
An MCP server exposes the same read surface as 7 tools for AI agents (Claude, Cursor, custom). One-line registration in ~/.claude/mcp.json. Tools include concept.get, concept.search, content.list, domain.overview, age_band.overview, health, contract.get.
Weekly gzipped JSONL dumps of every gold-tier concept, content item, edge, and source registered. Licensed CC-BY-SA 4.0 with attribution preserved down to the per-source row. Published at s3://koydo-meridian-public/.
Every source in the Meridian registry is CC-BY, CC-BY-SA, CC0, or public-domain. We do not ingest proprietary frameworks (IB, Core Knowledge, Cambridge, Pearson). We build our own taxonomy and cite upstream on every page.
Meridian powers the full Koydo learning stack — from early-years flashcards to graduate research prompts — and is openly accessible to partners, researchers, and AI agents.