For RAG admins

Turn your learning content
into a knowledge source for your AI.

Manabu2 lessons are structured text, organized into paths, courses, sections and lessons. You can read them through the API and MCP and use them as a source for your own RAG pipeline. Built-in RAG in Manabu2 is being designed.

See API docs Talk to engineering

Common challenges when you build RAG

High-quality RAG needs high-quality source documents.

No good sources

Internal docs are scattered, stale, and unstructured. Just getting them usable in a RAG index eats huge prep cost.

Knowledge lives in heads

Experts' tacit knowledge isn't written down. When the expert leaves, the knowledge leaves.

Content goes stale

Docs age out fast. Keeping RAG sources current is expensive maintenance.

Connecting learning content to your RAG

What you can use today is structured content and the API and MCP to read it. Chunking, embedding and retrieval happen in your own RAG stack.

Structured from the start

Content follows a path → course → section → lesson hierarchy, with one topic per lesson. That makes lessons easy to split on and easy to cite as sources.

Readable through the API and MCP

Apps and API keys with the read:content scope can fetch lesson bodies, materials and quiz questions. The search API and the MCP search_content tool find courses and paths.

Visibility carries over

An app token reads only what the signed-in user may see, and private courses stay inside your organization. Once you copy content into your own index, protecting it is up to your stack.

Built-in RAG is being designed

We're designing lesson-text indexing, answers with citations and a RAG management console inside Manabu2. They aren't available yet.

// Example: Manabu2 as a source for your own RAG
Manabu2 learning content
Paths, courses, sections and lessons
↓ Manabu2 API (read:content)
Your sync job & chunking
Split per lesson, re-fetch when content changes
↓ Embedding
Embedding (your choice of model)
e.g. Azure OpenAI / OpenAI
↓ Upsert
Your vector store
e.g. Azure AI Search / Pinecone / Weaviate
↓ Similarity search
Your internal AI assistant
"What's our company's process for X?" — answered
// Fetching a lesson body (read:content scope)
GET https://api.manabu2.com/api/v1/lessons/{lessonId}
Authorization: Bearer {access_token}

Ways to use learning content as a source

Internal AI assistant

Build a Slack/Teams bot on your own RAG stack that answers "what's the process for X?" or "what's our compliance rule?" from your organization's learning content.

Read straight from an AI assistant

Connect an MCP-capable AI assistant to Manabu2 and it can search courses and paths and read lesson bodies directly.

Course and path search

Manabu2's search finds courses and paths by their titles and descriptions. Clear titles and specific descriptions decide how findable your content is.

See other roles: For learners For HR For executives For curriculum managers For consultants

Make learning content the foundation for your AI

See the developer docs for how to use the API and authenticate.

See API docs Talk to engineering