Learning Path

Not sure where to start? This page maps out a clear progression through the CiniterFlow documentation. Each level builds on the previous one, so you develop skills in the right order.

How to Use This Page

• Go in order — each level assumes you've completed the previous one
• Don't skip the skill guides — Prompt Engineering and Model Selection will save you hours of trial and error
• Build every project — reading isn't enough; the hands-on practice is where the learning happens

---

Level 1: Foundations

Go from zero to understanding all 3 builders. No prior AI experience needed.

Time: ~3 hours

Step	Guide	What You'll Learn	Time
1.1	Sign Up & Onboarding	Create your account, get API keys, add credentials	10 min
1.2	Your First AI Assistant	Build a chatbot with the Assistant builder	20 min
1.3	Understanding the Dashboard	Navigate the CiniterFlow interface	10 min
1.4	Working with Chatflows	Build workflows with the Chatflow builder	25 min
1.5	Mastering Agentflow	All 14 node types, flow state, common patterns	30 min
1.6	Adding Knowledge	Document Stores, RAG, chunking, embeddings, vector stores	25 min
1.7	Connecting Tools & APIs	Built-in tools, custom tools, MCP, OpenAPI	20 min
1.8	Embedding Your Chatbot	Chat widget on your website	15 min
1.9	Using the API	Prediction API, SDKs, streaming, sessions	20 min
1.10	Variables, Memory & State	Variables, memory types, flow state	20 min

✅ Milestone: You can build a basic Assistant, Chatflow, or Agentflow with knowledge and tools, embed it on a website, and call it via API.

---

Level 2: Core Skills

Master the essential techniques every CiniterFlow builder needs. Two skill guides followed by four hands-on projects.

Time: ~4 hours | Prerequisites: Level 1 complete

Skill Guides

Read these before building — they'll make every project that follows significantly better.

Step	Guide	What You'll Learn
2.1	Prompt Engineering Masterclass	5-step prompt framework, few-shot examples, guardrails, CiniterFlow variables, model-specific tips
2.2	Model Selection & Costs	Choosing the right model per task, cost optimization, tiered usage strategy

Projects

Step	Project	What You'll Build	Difficulty	Key Skills
2.3	Basic RAG	Q&A agent that searches your documents	⭐ Beginner	Agent node + Document Store in Agentflow
2.4	FAQ Chatbot	FAQ bot embedded on a website	⭐ Beginner	End-to-end: Document Store → upsert → Agent → embed
2.5	Interacting with API	Event management agent with full CRUD	⭐⭐ Intermediate	Request tools (GET/POST/PUT/DELETE), OpenAPI Toolkit, HTTP node
2.6	Structured Output	Batch event creator from natural language	⭐⭐ Intermediate	JSON Structured Output, Iteration node, HTTP node, Flow State

✅ Milestone: You can build RAG agents, connect to REST APIs, and extract structured data from AI responses.

---

Level 3: Intermediate Patterns

Build multi-step, self-correcting, and multi-agent systems.

Time: ~6 hours | Prerequisites: Level 2 complete

Step	Project	What You'll Build	Difficulty	Key Skills
3.1	Agentic RAG	Self-correcting RAG with query regeneration	⭐⭐ Intermediate	Condition Agent, Loop node, Flow State updates, multi-branch flows
3.2	Tools & MCP	Agent with custom tools + MCP connections	⭐⭐ Intermediate	Custom Tool creation, MCP (Stdio + Streamable HTTP), GitHub MCP
3.3	Human in the Loop	Email reply agent with approval workflow	⭐⭐ Intermediate	Human Input node, Form Input, Loop for revision, Gmail tool
3.4	Customer Support	Triage system routing to specialized agents	⭐⭐ Intermediate	Condition Agent routing, single vs multi-agent design, prompting strategy
3.5	SQL Agent	Natural language to SQL with self-correction	⭐⭐ Intermediate	Custom Function, JSON output, dual validation loops
3.6	Agent as Tool	Parent agent delegating to child agentflows	⭐⭐ Intermediate	Agent As Tool, combining RAG + tool delegation
3.7	Multi-Document Q&A	Query across multiple financial documents	⭐⭐ Intermediate	Metadata filtering, Retriever Tools, Tool Agent
3.8	Web Scrape Q&A	Scrape websites and answer questions	⭐⭐ Intermediate	Cheerio/Playwright, upsertion, web crawling

✅ Milestone: You can build self-correcting flows, multi-agent routing, human-in-the-loop workflows, and SQL-powered agents.

---

Level 4: Advanced & Production

Build complex, production-ready AI systems for real business needs.

Time: ~8 hours | Prerequisites: Level 3 complete

Step	Project	What You'll Build	Difficulty	Key Skills
4.1	Supervisor & Workers	Write-review-refine pipeline	⭐⭐⭐ Advanced	Supervisor pattern, LLM routing, bidirectional loops
4.2	Deep Research Agent	Multi-agent research with dynamic subagents	⭐⭐⭐ Advanced	Iteration + Agent spawning, Planner→SubAgents→Writer→Condition loop
4.3	Email Auto-Responder	Draft replies with human approval	⭐⭐ Intermediate	Form Input, Agent + web search, Human Input, Loop, Gmail
4.4	Content Generator	Collaborative content creation pipeline	⭐⭐⭐ Advanced	Supervisor-Worker applied to content, Writer + Reviewer agents
4.5	Lead Outreach Agent	Research prospects, personalized emails, batch send	⭐⭐⭐ Advanced	Form, JSON parsing, Iteration, Agent research, Human Input, batch send
4.6	Meeting Scheduler	Calendar management via natural language	⭐⭐⭐ Advanced	Google Calendar OAuth, Gmail, Human Input on write operations
4.7	WhatsApp & Messaging Bot	Deploy agents on WhatsApp, Slack, Teams	⭐⭐⭐ Advanced	Prediction API, webhook middleware, sessionId mapping

✅ Milestone: You can build production-grade multi-agent systems, business automation pipelines, and deploy across messaging channels.

---

Additional Resources

These pages aren't part of the linear path but are essential references you'll use throughout your journey.

Chatflow-Specific Tutorials

If you're working with the Chatflow builder specifically, these tutorials cover Chatflow-native approaches:

• Interacting with API (Chatflow) — OpenAPI Chain and Toolkit in Chatflows
• API-Powered Assistant — Summary of all API integration approaches
• SQL Q&A (Chatflow) — SQL chatbot using Chatflow chains
• Calling Children Flows — Turn chatflows into reusable tools
• Calling Webhooks — Trigger external workflows via webhooks
• Pipedream MCP — Connect to 3,000+ APIs via Pipedream
• Upserting Data — Deep dive into the RAG data pipeline

Platform Reference

Deep-dive documentation for every CiniterFlow feature:

• Agentflow Node Reference — All 14 node types in detail
• Document Stores — Knowledge base management
• Prediction API — API and SDK reference
• Streaming — Real-time token streaming
• Embed Widget — Chat widget customization
• Uploads — Image, audio, and file processing
• Variables — Reusable values across flows
• Evaluations — Test and measure agent performance
• Marketplace & Templates — Pre-built flows

Analytics

• Arize | LangWatch | Langfuse | Opik | Phoenix

Integrations

• LangChain — Agents, chains, chat models, document loaders, embeddings, memory, tools, vector stores, and more
• LlamaIndex — Agents, chat models, embeddings, engines, tools, vector stores
• LiteLLM — Universal LLM proxy
• Utilities — Custom JS, If/Else, Set/Get Variable, Sticky Note
• External Integrations — Zapier, Open WebUI, Streamlit

API Reference

• Full API Reference — All 14 API endpoints