Family Learning Assistant

A study plan that starts from one sentence. The agent handles orchestration; the rules stay in code.

Independent projectLivePythonLangGraphReact
01

The problem

When you use AI to learn something, the friction is rarely "I can't get an answer". It's that the material lives in PDFs and screenshots you can't revise from, that what you learned today is unfindable in three days, and that every tool needs the context explained again.

So the product's main line is not conversation — it is turning material into vocabularies and tasks. Chat is the entry point; the output is structured learning material.

02

The live product

Login page of the product: product statement and an agent conversation example on the left, sign-in form on the right
yidongxun.com login page (screenshot, Nov 2026). The left panel states the main path: one sentence → agent configuration preview → confirm → write → dedicated practice page.
Mobile login page
Mobile, 390px
Shipped capabilities
  • Conversational intent parsing and tool orchestration
  • Study-material parsing and term extraction
  • Vocabulary creation and organisation
  • Dictation task generation
  • A dedicated practice page for dictation
  • Human confirmation before create / update / delete
  • Sign-in and per-user data isolation

Derived from the login page copy and the product's actual path; details behind the login are as the product has them.

03

Core flow

1 Material→2 Parse→3 Confirm→4 Vocabulary→5 Task→6 Practice

Step 3 is deliberate: terms extracted by the model only enter the store after a person confirms them, so wrong content never becomes study material.

04

Agent architecture

Request path
React client
Chat & confirm
FastAPI
Auth · session · SSE
Main agent
Intent · pick tool
Subgraph
Fixed-order steps
Execution & data
Tool registry
Schema · routing
Executor
Validate · retry · timeout
Service
Perms · txn · state
MySQL
Business data · ckpt

Schematic, to show layering and call direction; module names and details follow the code.

05

Engineering decisions

The agent never touches the database

The model writes no SQL and knows no table names. It can only choose which tool to call; the tool goes through the service layer.

The service layer is the only entry point

Business rules, permissions, transactions and state checks cannot be written or verified in a prompt. The backend is split into 11 domain modules, cross-module repository access is forbidden, and a dedicated user module owns authentication and data isolation.

Tool calling and workflow are separate jobs

Fixed order becomes edges; only semantics-dependent branches go to the model. One main agent, internal subgraphs, an LLM that only understands intent and picks tools.

Checkpoints, confirmation and recovery

LangGraph checkpointers let a long task resume from the last trusted state instead of restarting the whole chain; confirmation points show the result first, then execute, with idempotent side effects.

06

Stack

PythonFastAPILangGraphReactTypeScriptMySQLSSELangSmithDeepSeek
07

From zero to production

Moving from a Java backend to the AI agent stack, I designed and shipped this product alone: product definition, architecture, full stack and deployment. Everything that went wrong on the way — tool boundaries, context growth, state recovery, cost and latency — became 28 articles.