DEV.co
RAG Frameworks · didilili

ai-agents-from-zero

ai-agents-from-zero is a Chinese-language open-source tutorial and learning path for building AI agents and large language model applications. It covers frameworks like LangChain and LangGraph, low-code platforms like Dify and Coze, and includes executable projects, interview prep, and enterprise deployment guidance.

Source: GitHub — github.com/didilili/ai-agents-from-zero
2.6k
GitHub stars
353
Forks
Python
Primary language
MIT
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
Repositorydidilili/ai-agents-from-zero
Ownerdidilili
Primary languagePython
LicenseMIT — OSI-approved
Stars2.6k
Forks353
Open issues12
Latest releaseUnknown
Last updated2026-06-23
Sourcehttps://github.com/didilili/ai-agents-from-zero

What ai-agents-from-zero is

A Python-focused educational repository combining LLM fundamentals, prompt engineering, Tool Calling, RAG (vector + sparse + reranking), LangGraph workflows, MCP protocol integration, and multi-agent patterns. Includes two completed end-to-end projects: NL2SQL e-commerce Q&A (LangGraph + MySQL) and multi-agent deep research (DeepAgents framework).

Quickstart

Get the ai-agents-from-zero source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/didilili/ai-agents-from-zero.gitcd ai-agents-from-zero# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Learning AI Agent architecture and LangGraph workflows

Comprehensive curriculum starting from LLM basics through production RAG/Agent patterns, with runnable code examples and project templates.

Building production RAG and agent systems in Python

Practical focus on enterprise-grade retrieval (vector + sparse + reranking), intent routing, human handoff, and multi-path recall patterns with working code.

Interview preparation for AI/LLM application engineer roles

Dedicated interview question bank aligned to job descriptions and training curricula, covering conceptual, code, and system design domains.

Implementation considerations

  • All projects use external LLM APIs (OpenAI, cloud-hosted models) or local deployments (Ollama, Xinference); plan for cost and latency accordingly.
  • Tutorials reference Coze (Tencent platform), Dify, and other low-code tools—adoption depends on infrastructure availability and internal tooling strategy.
  • Multi-step RAG pipelines (vector + sparse + reranking) add operational complexity; evaluate RAGAS evaluation metrics and monitoring requirements upfront.
  • Projects use PostgreSQL, Qdrant, Elasticsearch, Neo4j for different retrieval patterns—data pipeline setup is non-trivial.
  • LangGraph and MCP are production-grade but younger than LangChain; community support and breaking changes should be monitored.

When to avoid it — and what to weigh

  • Non-Python tech stacks (Java, C#, Go) — Content is almost entirely Python-centric (LangChain, LangGraph). Java equivalents (langchain4j) are explicitly not covered.
  • Non-English speakers requiring English-only resources — Repository and tutorial are primarily in Chinese. Not suitable for English-only teams without translation capacity.
  • Proprietary/closed-source enterprise deployment requirements — While MIT-licensed, content focuses on open-source and some cloud-platform (Coze, Dify) deployment. May not align with strict on-premise or closed-model mandates.
  • Beginners with zero coding experience — Assumes Python familiarity and software engineering fundamentals. Target is engineers transitioning to AI development, not absolute beginners.

License & commercial use

Licensed under MIT License. Permissive open-source license allows commercial use, modification, and distribution with minimal restrictions. Requires only attribution and preservation of license notice.

MIT License explicitly permits commercial use of the educational content and code examples. However, projects in the tutorial may depend on external APIs (OpenAI, Coze, Dify) or third-party libraries with their own license restrictions. Always review transitive dependencies and API terms of service before deploying to production. No warranty or liability guarantees implied.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationStrong
License clarityClear
Deployment complexityModerate
DEV.co fitStrong
Assessment confidenceHigh
Security considerations

No explicit security audit or vulnerability disclosure process documented. Standard considerations apply: secure storage of API keys (use environment variables, secrets management), input validation for user prompts (injection risk), rate limiting on external API calls, and PII handling in chat logs. MCP server implementations should enforce authentication and authorization. No statement on data retention or compliance (GDPR, etc.).

Alternatives to consider

LangChain official documentation + LangGraph tutorials

Official source, actively maintained by LangChain team; narrower scope (no full project arc or interview prep) but authoritative for framework details.

Dify + Coze native documentation + tutorials

Vendor-native resources for low-code agent building; less emphasis on custom Python development or multi-framework comparison.

DeepLearning.AI short courses (LangChain, RAG, Agents)

English-language, shorter-form structured courses; less comprehensive enterprise project coverage but higher production value and live instructor support.

Software development agency

Build on ai-agents-from-zero with DEV.co software developers

Explore the tutorial repository, run the executable projects, and prepare for AI engineer interviews—all in one integrated curriculum.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

Related on DEV.co

Explore the category and the services that help you build with it.

ai-agents-from-zero FAQ

Is this suitable for production use, or just learning?
Both. The tutorial spans from conceptual foundations to enterprise-grade patterns (RAG evaluation, multi-path recall, monitoring). Two end-to-end projects (e-commerce Q&A and multi-agent research) are designed to be deployable. However, you will need to adapt examples to your data, infrastructure, and API choices.
Do I need to know LangChain before starting?
No formal prerequisite, but Python familiarity and basic software engineering knowledge (APIs, databases, deployment) are assumed. The tutorial covers LangChain from fundamentals; it is designed as a learning path, not a reference manual.
Are the projects runnable out of the box?
Mostly yes, with caveats: you will need API keys (OpenAI, etc.), a vector DB and other infrastructure services, and some environment setup. The tutorial emphasizes 'runnable' as a standard, and linked source repos (shopkeeper-agent, deepsearch-agents) are standalone GitHub projects. Common setup issues are documented in the FAQ section.
Is this only for Chinese speakers?
The tutorial and README are in Chinese. Code examples and framework names are English. If your team is non-Chinese, expect to use translation tools or allocate time for language review. The conceptual frameworks and project structures are language-agnostic.

Software developers & web developers for hire

Adopting ai-agents-from-zero is usually one piece of a larger software development effort. As a software development agency, DEV.co provides software development services and web development expertise — pairing senior software developers and web developers with your team to design, build, and operate rag frameworks software in production.

Ready to build AI agents? Start your learning path.

Explore the tutorial repository, run the executable projects, and prepare for AI engineer interviews—all in one integrated curriculum.