From your first widgetto fully-offline AI.
Three books, one reading path. Build a real Flutter app, give it production AI, then run that AI privately on the device — with every line of code verified for June 2026.
- 3 books
- 850+ pages
- Verified June 2026
- Flutter 3.44 · Dart 3.12
A deliberate path — not a pile of tutorials.
Each book hands off to the next. You build one real app — Habito — and carry it from a clean architecture, to production AI features, to AI that runs entirely on the device.
The Complete Flutter Blueprint
From Visual Prototypes to Seamless Store Distribution
The end-to-end craft of shipping a real Flutter app — from a design file to a polished build in the store.
Building AI-Powered Flutter Apps
From First Token to Production
Add real AI to a Flutter app and take it all the way to production — streaming chat, RAG, tools and agents, on-device and hybrid, then the unglamorous craft that keeps it shipping: evals, cost, and safety.
On-Device AI for Flutter
Private, Offline, and Free at Inference
Run real AI on the phone — no servers, no API bills, nothing leaves the device.
The AI Backend in Dart
The Server Every AI App Needs
Your AI app's provider key is sitting inside the app right now — and a stranger can drain it in minutes.
Ship a Real AI SaaS
From Idea to Paying Customers
A working AI demo is worth $0; a working AI business is worth everything — and the gap between them is this book.
Production AI Agents
Build, Trust, and Ship Agents That Do Real Work
A chatbot tells you the refund policy; an agent issues the refund — and that one word, "issues," changes everything about how you build and trust it.
Read straight through, or drop into the book you need — every chapter tells you up front what it assumes.
Three books. Read in any order.
Premium, fixed-layout PDFs that render code exactly the way you wrote it — built around one app you actually ship.
The Complete Flutter Blueprint
From Visual Prototypes to Seamless Store Distribution
The end-to-end craft of shipping a real Flutter app — from a design file to a polished build in the store. Twenty-nine chapters and a full capstone, with a locked, modern stack and nothing left hand-wavy.
- Translate a design into a clean widget tree, the right way
- A locked production stack: Riverpod, Drift, dio, go_router
- Build Habito — a real habit tracker — end to end
- Test, profile, and ship to both stores with confidence
Building AI-Powered Flutter Apps
From First Token to Production
Add real AI to a Flutter app and take it all the way to production — streaming chat, RAG, tools and agents, on-device and hybrid, then the unglamorous craft that keeps it shipping: evals, cost, and safety.
- Your first real token, then a streaming chat UI that doesn't jank
- Embeddings, RAG, and multimodal that actually ground answers
- Agents, tools, and MCP — with guardrails that hold
- Evals, caching, cost, and safety: the production half nobody writes about
On-Device AI for Flutter
Private, Offline, and Free at Inference
Run real AI on the phone — no servers, no API bills, nothing leaves the device. Local LLMs, offline RAG, vision and speech, performance under heat and memory pressure, and a capstone that works in airplane mode.
- Run Gemma on a phone with flutter_gemma and LiteRT
- Fully-offline RAG — a private knowledge base that never phones home
- Survive real hardware: memory, battery, and thermal throttling
- Ship a fully-offline feature that works on a plane, at zero cost
The AI Backend in Dart
The Server Every AI App Needs
Your AI app's provider key is sitting inside the app right now — and a stranger can drain it in minutes. This book moves the intelligence server-side in Dart: a streaming key-proxy, RAG over private data, auth, metered billing, background jobs, and deploy — the backend every real AI app needs, built end to end with Dart Frog.
- Move the provider key server-side behind a clean proxy — the vulnerability, gone
- Stream tokens through your server, and give the model tools next to your data
- Postgres + pgvector RAG, per-user auth, metered billing, and background jobs
- Containerize and deploy to Fly.io / Cloud Run — plus an honest Serverpod head-to-head
Ship a Real AI SaaS
From Idea to Paying Customers
A working AI demo is worth $0; a working AI business is worth everything — and the gap between them is this book. Build Cortex, a real AI SaaS, end to end on Next.js 16 and the Vercel AI SDK: streaming chat, RAG over your users' documents, tools and generative UI, multi-tenant auth, Stripe subscriptions, usage metering, and a growth loop — from create-next-app to a product that could take its first paying customer.
- Stream a chat, give the model tools, and render generative UI — the AI product core
- Multi-tenant data, embeddings, and a production RAG pipeline over your users' documents
- Auth and workspaces, Stripe subscriptions, usage metering, and safety — the business layer
- Deploy on Vercel, watch cost and evals, and wire the marketing site + growth loop that sells it
Production AI Agents
Build, Trust, and Ship Agents That Do Real Work
A chatbot tells you the refund policy; an agent issues the refund — and that one word, "issues," changes everything about how you build and trust it. Give a model tools and a loop, connect it to the world with MCP, make it survive crashes, then earn the right to ship it: human approval on dangerous actions, guardrails against prompt injection, evals that prove it works, tracing that shows what it did, and cost control that keeps it viable. Every API verified against a compiling project.
- Build a tool-using agent in twenty lines, then master the tool craft that decides whether it succeeds
- Connect to the world with MCP — consume any server, and publish your own tools as one
- Steer the loop, split work across subagents, and make long-running agents survive crashes and deploys
- Earn trust: human approval, prompt-injection defense, evals, tracing, and cost per completed task
The AI-Flutter Stack
The complete vertical: build a real Flutter app, give it production AI, then run that AI privately on the device. One reading path, first widget to airplane mode.
- The Complete Flutter Blueprint· 523 pp
- Building AI-Powered Flutter Apps· 186 pp
- On-Device AI for Flutter· 143 pp
- The AI Backend in Dart· 177 pp
- Ship a Real AI SaaS· 157 pp
- Production AI Agents· 139 pp
Flutter Field Kit
Cheat-Sheets & Interview Prep
The reference you keep in a second tab — and the one you cram the night before an interview. Built on the parts of Flutter that don’t go stale, so it stays correct: every snippet analyzer-checked, no version pins anywhere.
- Eleven cheat-sheets built as lookup tables — find the row, take the answer, get back to work
- 112 interview questions graded junior → senior, each with the signal the interviewer is really after
- Six mobile system-design prompts with a frame that works when your mind goes blank
Written like production code — reviewed like it too.
In a market flooded with AI-generated filler, these are built for developers who have to make it work on Monday.
Verified for June 2026
Every package version, API name, and signature fact-checked against the live docs — not a model's memory. The AI stack moves fast; this keeps up.
Runnable, never hand-wavy
Code that compiles against the real SDKs — firebase_ai, flutter_gemma, LiteRT — and is adversarially reviewed for correctness.
One app, three books
You build Habito once and carry it the whole way — architecture, cloud AI, then fully-offline AI. The thread never drops.
Diagrams that explain
30+ hand-built diagrams — runtime maps, RAG loops, routing trees — so the hard ideas land before the code does.
Honestly offline
On-device means on-device: airplane-mode-provable, zero API cost. The book tells you exactly where it works and where it doesn't.
Honest about limits
Where small models fail, where RAG beats fine-tuning, where the cloud still wins. No hype — the trade-offs you actually ship against.
A reading experience made for developers.
Fixed-layout PDFs that never mangle a code block, plus a free in-browser reader for samples. No tiny re-flowed snippets, no broken indentation — the way technical books should read.
- Code that renders exactly as written
- Built for long, dark-mode reading
- Reads cleanly on any screen
The Case for On-Device AI
The model that answers your user’s question doesn’t have to live in a data center. It can live in their pocket — running on the phone they’re already holding, answering with no network, no API bill, and nothing leaving the device.
That single shift changes the economics and the ethics of a feature at once. Private by construction. The only honest test is the simplest one: turn on airplane mode and see if it still works.
Most AI-development content is either toy demos or already out of date. NIM Books exists to fix that: deeply practical, ruthlessly current, and anchored in Flutter — the craft of shipping real apps with AI that actually holds up in production.
Every book is built the same way the code is: researched against live sources, written for a developer who has to make it work, and reviewed until it’s right.
Start with a free chapter.
Grab a free chapter— “From Design to Widgets” — instantly. Then get the full books whenever you’re ready.
No spam. The full books are available now — this is just a taster.