# How Makers Build > The 22 decisions a maker makes building a web SaaS or an AI app — framework, database, auth, payments, models, hosting — which tool fits which situation, what real makers chose (every usage linked to its source: the maker's own words, a customer story, a subprocessor list, the product's website, or its public code), and which tools AI coding agents can work with (llms.txt, official MCP server, CLI). Tools are described by situation, never ranked. Counts are what can be observed in makers' live products, not market share. ## Starter stacks - [Ship an AI SaaS this weekend](https://howmakersbuild.com/stacks/ai-saas-weekend): You're one person, you write TypeScript, and the AI feature is the product. You want real users signing in and paying by Monday, and you'd rather pay a little per month than run servers. - [Sell worldwide without handling tax](https://howmakersbuild.com/stacks/sell-worldwide): You'll have customers in dozens of countries from the first week, you're not a company that can register for VAT everywhere, and you want someone else to be the seller of record. - [AI app that bills by usage](https://howmakersbuild.com/stacks/ai-usage-billing): Every request costs you model tokens, so a flat subscription either loses money on heavy users or prices out light ones. You want to charge per use, credits or tiers with limits — and see what each user costs you. - [Build it with an AI app builder](https://howmakersbuild.com/stacks/ai-app-builder): You'd rather describe the product than write it. You'll build in an AI app builder, and you want the pieces it wires up for you — so the result works on day one and a developer can take it over later. - [Chat with your docs (RAG)](https://howmakersbuild.com/stacks/rag-app): Your product answers questions from a pile of documents — help docs, contracts, a knowledge base. You need to ingest files, search them by meaning, and stream answers that cite their sources. - [B2B SaaS that sells to companies](https://howmakersbuild.com/stacks/b2b-saas): Your customers are teams inside companies. Sooner or later an IT department asks for SSO, audit trails and invoices, and a lost email or an outage costs you a contract. - [Python AI backend with a web front end](https://howmakersbuild.com/stacks/python-ai-backend): The AI part of your product lives in Python — models, data work, the libraries researchers publish first — and the front end is a separate web app talking to it. - [$0 until you have users](https://howmakersbuild.com/stacks/free-tier-launch): You're validating an idea and don't want a single monthly bill until real people use it. Every pick here has a free tier a small product can run on in production. - [An AI agent that does real work](https://howmakersbuild.com/stacks/ai-agent-product): Your product is an agent — it plans, calls tools and APIs, waits on people or other systems, and runs for minutes or hours on a user's behalf. Reliability and cost matter more than a chat box. - [The one-person monolith](https://howmakersbuild.com/stacks/monolith): You want one codebase, one deploy and batteries included — models, admin, jobs and email built in — and you'd rather write product than wire services together. - [Edge-first on Cloudflare](https://howmakersbuild.com/stacks/cloudflare-edge): You want your app and data close to users everywhere, a bill that stays near zero at low traffic, and no egress fees — and you're happy to build on one platform's primitives. ## Foundation - [Frontend Framework](https://howmakersbuild.com/tasks/frontend-framework): Which full-stack framework, and so which UI library under it. The React frameworks (Next.js, React Router, TanStack Start) get the largest pool of components, examples and AI-agent familiarity; SvelteKit and Nuxt trade that for Svelte's or Vue's style; Astro is content-first and ships almost no JavaScript, but isn't built for app-heavy screens. A bare UI library on Vite is still enough for a single-page app behind a login. - [React](https://howmakersbuild.com/tools/react) (UI library (with Vite)): Single-page apps built from React components, with the largest ecosystem of component libraries, examples and AI-generated code. Chosen by 200 makers' products we track. - [Next.js](https://howmakersbuild.com/tools/nextjs) (React full-stack): React apps that want server rendering, API routes and the largest set of examples, templates and SaaS starters. Chosen by 1510 makers' products we track. - [React Router](https://howmakersbuild.com/tools/react-router) (React full-stack): React apps built around web standards — loaders, actions and forms — in the lineage of Remix. - [TanStack Start](https://howmakersbuild.com/tools/tanstack-start) (React full-stack): Type-safe routing and server functions for React or Solid, built on TanStack Router. - [SvelteKit](https://howmakersbuild.com/tools/sveltekit) (Svelte or Vue full-stack): Apps written in Svelte, with less boilerplate and small client bundles. - [Nuxt](https://howmakersbuild.com/tools/nuxt) (Svelte or Vue full-stack): Vue apps, with a large module ecosystem for auth, content, images and SEO. Chosen by 30 makers' products we track. - [Astro](https://howmakersbuild.com/tools/astro) (Content-first): Marketing sites, docs and blogs that ship minimal JavaScript, with interactive islands from any UI library. Chosen by 212 makers' products we track. - [Backend Framework](https://howmakersbuild.com/tasks/backend-framework): Minimal router or batteries included. Minimal frameworks handle HTTP and leave the database, auth and admin to you; batteries-included frameworks give you ORM, migrations, auth and background jobs with one set of conventions, at the cost of learning them. - [Hono](https://howmakersbuild.com/tools/hono) (Minimal router): Small TypeScript APIs that run on Node, Bun, Deno or edge platforms like Cloudflare Workers. Chosen by 3 makers' products we track. - [Express](https://howmakersbuild.com/tools/express) (Minimal router): Node.js servers where you want a widely documented default and a huge middleware catalog. Chosen by 3 makers' products we track. - [Fastify](https://howmakersbuild.com/tools/fastify) (Minimal router): Node.js APIs that want low overhead, JSON-schema request validation and a structured plugin system. - [FastAPI](https://howmakersbuild.com/tools/fastapi) (Minimal router): Python APIs with request validation and interactive docs generated from type hints. Chosen by 29 makers' products we track. - [Django](https://howmakersbuild.com/tools/django) (Batteries included): Python apps that want an ORM, migrations, auth and an admin panel on day one. Chosen by 37 makers' products we track. - [Ruby on Rails](https://howmakersbuild.com/tools/rails) (Batteries included): Full-stack Ruby apps where one developer builds models, views, jobs and emails with strong conventions. Chosen by 24 makers' products we track. - [Laravel](https://howmakersbuild.com/tools/laravel) (Batteries included): PHP apps with ORM, queues, auth scaffolding and an optional paid ecosystem for hosting and admin panels. Chosen by 73 makers' products we track. - [NestJS](https://howmakersbuild.com/tools/nestjs) (Structured TypeScript): Larger TypeScript backends organized into modules, with dependency injection and decorators. Chosen by 21 makers' products we track. - [Database](https://howmakersbuild.com/tasks/database): Where the data lives: a whole backend platform (database, auth and storage in one project), a managed database you reach from your own code, or the raw engine on your own server. Postgres is the default unless you have a reason to pick otherwise; analytics and graph databases sit next to it for workloads it handles poorly, not instead of it. How your code talks to it is its own decision (Database Access & ORMs). - [Supabase](https://howmakersbuild.com/tools/supabase) (Backend platform): Postgres plus auth, storage and realtime in one project. Chosen by 564 makers' products we track. - [Convex](https://howmakersbuild.com/tools/convex) (Backend platform): Realtime, collaborative apps written entirely in TypeScript. Chosen by 50 makers' products we track. - [Firebase](https://howmakersbuild.com/tools/firebase) (Backend platform): Mobile apps that need offline sync and client-side SDKs. Chosen by 182 makers' products we track. - [Neon](https://howmakersbuild.com/tools/neon) (Managed database): Serverless Postgres that scales to zero, with a branch per preview deploy. Chosen by 95 makers' products we track. - [PlanetScale](https://howmakersbuild.com/tools/planetscale) (Managed database): Apps that expect heavy scale and want zero-downtime schema changes. Chosen by 29 makers' products we track. - [Turso](https://howmakersbuild.com/tools/turso) (Managed database): SQLite at the edge, or one database per customer. Chosen by 13 makers' products we track. - [MongoDB](https://howmakersbuild.com/tools/mongodb) (Managed database): Document-shaped data and teams already fluent in Mongo. Chosen by 131 makers' products we track. - [PostgreSQL](https://howmakersbuild.com/tools/postgresql) (Run it yourself): The engine behind most of the hosted options above, run on your own server or container for full control and a flat cost. Chosen by 74 makers' products we track. - [Database Access & ORMs](https://howmakersbuild.com/tasks/orm): Schema-first ORM, SQL-shaped query builder, or the platform's own SDK. A schema-first ORM generates types and migrations from one schema file and hides the SQL; a SQL-shaped ORM or query builder keeps queries close to SQL with full types and a smaller runtime; a backend platform (Supabase, Convex, Firebase) comes with its own client, and adding an ORM on top is often unnecessary. - [Prisma](https://howmakersbuild.com/tools/prisma) (Schema-first ORM): A schema-first ORM with generated types and migrations. Chosen by 37 makers' products we track. - [Drizzle ORM](https://howmakersbuild.com/tools/drizzle) (SQL-shaped ORM): A thin, SQL-like TypeScript ORM that runs well on serverless and edge runtimes. Chosen by 10 makers' products we track. - [Kysely](https://howmakersbuild.com/tools/kysely) (Query builder): A type-safe SQL query builder for TypeScript, with no schema file or code generation step of its own. - [SQLAlchemy](https://howmakersbuild.com/tools/sqlalchemy) (SQL-shaped ORM): Python backends (FastAPI, Flask, scripts) that want an ORM or a SQL toolkit over any relational database. - [Authentication](https://howmakersbuild.com/tasks/authentication): Hosted service or library in your own app. Hosted services give you UI, MFA and security updates for a monthly bill and some lock-in; libraries keep users in your own database for free, and you own the upkeep. - [Clerk](https://howmakersbuild.com/tools/clerk) (Hosted service): Dropping prebuilt sign-in, profile and organization UI into a React or Next.js app. Chosen by 307 makers' products we track. - [Supabase](https://howmakersbuild.com/tools/supabase) (Bundled with your backend): Apps already on Supabase — auth is part of the same project as the database. Chosen by 13 makers' products we track. - [Firebase](https://howmakersbuild.com/tools/firebase) (Bundled with your backend): Mobile and client-heavy apps already using Firebase. Chosen by 4 makers' products we track. - [Better Auth](https://howmakersbuild.com/tools/better-auth) (Library in your app): TypeScript apps that want users and sessions in their own database, at no cost. Chosen by 6 makers' products we track. - [WorkOS](https://howmakersbuild.com/tools/workos) (Hosted service): B2B SaaS that will need enterprise SSO and directory sync. Chosen by 19 makers' products we track. - [Auth0](https://howmakersbuild.com/tools/auth0) (Hosted service): Complex identity needs and enterprise compliance. Chosen by 21 makers' products we track. - [File Storage](https://howmakersbuild.com/tasks/file-storage): Raw object storage or an upload service. Buckets (S3 and the S3-compatible clones) are cheap and portable but leave uploads, validation and access rules to you; upload services and platform-bundled storage hand you a working upload flow and pick the bucket for you. - [Amazon S3](https://howmakersbuild.com/tools/aws-s3) (Object storage): Apps already on AWS, or ones that need lifecycle rules, storage tiers and event triggers. - [Cloudflare R2](https://howmakersbuild.com/tools/cloudflare-r2) (Object storage): Files that get downloaded often — the S3 API without egress fees. Chosen by 4 makers' products we track. - [UploadThing](https://howmakersbuild.com/tools/uploadthing) (Upload service): TypeScript apps that want typed upload endpoints and drop-in upload components instead of wiring presigned URLs. - [Vercel Blob](https://howmakersbuild.com/tools/vercel-blob) (Bundled with your platform): Next.js apps hosted on Vercel that need simple uploads with no extra account. - [Supabase](https://howmakersbuild.com/tools/supabase) (Bundled with your platform): Apps already on Supabase — storage access follows the same row-level security policies as your data. Chosen by 3 makers' products we track. - [Background Jobs & Cron](https://howmakersbuild.com/tasks/background-jobs): Hosted job platform or a queue you run. Hosted platforms work on serverless hosting and give you retries, schedules and a dashboard, at the cost of a vendor in your critical path; a queue library is free and fully yours, but needs a long-running worker process and Redis or Postgres to hold the jobs. A durable workflow engine sits between: you run the workers, it keeps each workflow's state so long processes resume after failures. - [Trigger.dev](https://howmakersbuild.com/tools/trigger-dev) (Hosted job platform): Long-running TypeScript tasks — AI calls, media processing — that would time out in a serverless function. Chosen by 23 makers' products we track. - [Inngest](https://howmakersbuild.com/tools/inngest) (Hosted job platform): Event-driven, multi-step workflows where each step retries on its own and your code stays in your existing app. Chosen by 16 makers' products we track. - [Temporal](https://howmakersbuild.com/tools/temporal) (Durable workflow engine): Long-running, multi-step workflows — payments, onboarding, AI agent loops — whose state must survive crashes and deploys, in Go, Java, Python, TypeScript and more. Chosen by 29 makers' products we track. - [Upstash](https://howmakersbuild.com/tools/upstash) (Serverless queue): Simple scheduled calls and delayed HTTP jobs on serverless or edge hosting, via QStash. Chosen by 2 makers' products we track. - [BullMQ](https://howmakersbuild.com/tools/bullmq) (Self-run queue): Node apps on a long-running server that already have Redis — free and fully in your control. Chosen by 1 makers' products we track. - [Transactional Email](https://howmakersbuild.com/tasks/transactional-email): Transactional or marketing. Transactional services send one message the moment your app triggers it (a password reset, a receipt). Marketing/lifecycle tools send sequences and campaigns to lists. Most apps need the first on day one and the second later. - [Resend](https://howmakersbuild.com/tools/resend) (Transactional API): A modern developer API with email templates written as React components. Chosen by 944 makers' products we track. - [Postmark](https://howmakersbuild.com/tools/postmark) (Transactional API): Deliverability when the mail has to arrive within seconds. Chosen by 615 makers' products we track. - [Mailgun](https://howmakersbuild.com/tools/mailgun) (Transactional API): High volume and inbound email parsing and routing. Chosen by 49 makers' products we track. - [Amazon SES](https://howmakersbuild.com/tools/amazon-ses) (Raw infrastructure): Sending at high volume, where cost per email matters more than tooling. Chosen by 4 makers' products we track. - [SendGrid](https://howmakersbuild.com/tools/sendgrid) (Transactional + marketing): One vendor for both jobs at large scale. Chosen by 1061 makers' products we track. ## Money - [Payments](https://howmakersbuild.com/tasks/payments): Payment processor or merchant of record (MoR). With a processor you are the seller and must register for and remit sales tax/VAT yourself. An MoR resells your product, so tax and compliance are theirs, for a higher fee. - [Stripe](https://howmakersbuild.com/tools/stripe) (Payment processor): Full control over checkout and billing, lower fees than a merchant of record, and the widest integration ecosystem. Chosen by 891 makers' products we track. - [PayPal](https://howmakersbuild.com/tools/paypal) (Payment processor): Adding PayPal and Venmo as a checkout option for buyers who won't type in a card. Chosen by 101 makers' products we track. - [Polar](https://howmakersbuild.com/tools/polar) (Merchant of record): Developers selling SaaS, subscriptions, usage-based plans or license keys. Chosen by 50 makers' products we track. - [Lemon Squeezy](https://howmakersbuild.com/tools/lemon-squeezy) (Merchant of record): Selling software and digital downloads with minimal setup. Chosen by 77 makers' products we track. - [Paddle](https://howmakersbuild.com/tools/paddle) (Merchant of record): Established SaaS businesses that want a mature MoR. Chosen by 79 makers' products we track. - [Dodo Payments](https://howmakersbuild.com/tools/dodo-payments) (Merchant of record): AI and SaaS products with usage-based billing, selling globally. Chosen by 30 makers' products we track. ## AI in the product - [LLM API](https://howmakersbuild.com/tasks/llm-api): Which provider, and whether to go through a gateway. Going direct to a provider gets its newest models and features first and the lowest latency; a gateway gives you one key and one bill for hundreds of models, with fallbacks when a provider is down, for a small markup or a proxy you run. The code that calls the model — an AI SDK or agent framework — is its own decision. - [OpenAI API](https://howmakersbuild.com/tools/openai) (Model provider): The broadest model lineup — text, images, speech, embeddings — and the largest ecosystem. Chosen by 734 makers' products we track. - [Claude](https://howmakersbuild.com/tools/claude) (Model provider): Coding, long documents and agent-style tool use. Chosen by 251 makers' products we track. - [Gemini API](https://howmakersbuild.com/tools/gemini-api) (Model provider): Very long context and multimodal input, with a free tier to start. Chosen by 99 makers' products we track. - [Groq](https://howmakersbuild.com/tools/groq) (Open models, hosted): Very low-latency inference on open models, for real-time features. Chosen by 39 makers' products we track. - [Mistral AI](https://howmakersbuild.com/tools/mistral) (Model provider): A European provider whose API covers chat, coding, OCR and speech models. Chosen by 15 makers' products we track. - [DeepSeek](https://howmakersbuild.com/tools/deepseek) (Model provider): Low per-token prices on strong reasoning and coding models, with OpenAI- and Anthropic-format endpoints. Chosen by 11 makers' products we track. - [OpenRouter](https://howmakersbuild.com/tools/openrouter) (Gateway (many providers, one API)): Trying many models from many providers with one API key and one bill, without opening an account at each. Chosen by 61 makers' products we track. - [LiteLLM](https://howmakersbuild.com/tools/litellm) (Gateway (many providers, one API)): Running your own OpenAI-compatible proxy in front of your own provider keys, with retries, routing and per-key budgets. Chosen by 21 makers' products we track. - [AI SDK & Agent Framework](https://howmakersbuild.com/tasks/agent-framework): Broad framework or thin SDK. Broad frameworks come with integrations for loaders, vector stores and memory, but add layers of abstraction you have to learn and debug. Thin SDKs give you a typed agent loop close to the raw API and leave the rest to you. An agent harness goes further the other way: the whole loop — file access, shell, context management — comes ready-made, for agents that work inside a workspace. - [LangChain](https://howmakersbuild.com/tools/langchain) (Broad framework): The largest set of ready-made integrations for models, tools and data sources, in Python and JavaScript. Chosen by 65 makers' products we track. - [LlamaIndex](https://howmakersbuild.com/tools/llamaindex) (Broad framework): Agents that mostly retrieve and reason over your own documents. Chosen by 17 makers' products we track. - [Mastra](https://howmakersbuild.com/tools/mastra) (TypeScript framework): TypeScript and Next.js apps that want agents, workflows, memory and RAG without a Python service. Chosen by 24 makers' products we track. - [OpenAI Agents SDK](https://howmakersbuild.com/tools/openai-agents-sdk) (Thin SDK): A small agent loop with tools, handoffs, guardrails and tracing built in, from the OpenAI team. Chosen by 1 makers' products we track. - [Pydantic AI](https://howmakersbuild.com/tools/pydantic-ai) (Thin SDK): Python agents whose outputs must validate against typed schemas, across any model provider. - [AI SDK](https://howmakersbuild.com/tools/vercel-ai-sdk) (Thin SDK): TypeScript agents with tool calling, multi-step loops and streaming to the UI, on any model provider. Chosen by 4 makers' products we track. - [Claude Agent SDK](https://howmakersbuild.com/tools/claude-agent-sdk) (Agent harness): Agents that read and edit files, run shell commands and manage long context, built on the same harness as Claude Code. - [LangGraph](https://howmakersbuild.com/tools/langgraph) (Multi-agent orchestration): Stateful agents built as an explicit graph of steps, with checkpoints, retries, streaming and pauses for human approval, in Python or JavaScript. Chosen by 1 makers' products we track. - [Vector Database](https://howmakersbuild.com/tasks/vector-database): Inside the database you already run, or a separate vector store. pgvector keeps vectors next to your rows, with one backup and one query language, but tuning large indexes is on you. A dedicated vector store handles scale, filtering and hybrid search for you, at the cost of another service to sync data into. - [pgvector](https://howmakersbuild.com/tools/pgvector) (Inside your database): Apps already on Postgres that want vector search in the same database, joined with normal tables. - [Chroma](https://howmakersbuild.com/tools/chroma) (Embedded / local-first): Prototyping RAG on your laptop with a pip or npm install and no server to run. Chosen by 13 makers' products we track. - [Pinecone](https://howmakersbuild.com/tools/pinecone) (Managed vector store): A fully hosted index with nothing to operate, sized by usage. Chosen by 55 makers' products we track. - [Qdrant](https://howmakersbuild.com/tools/qdrant) (Open-source vector store): Heavy metadata filtering alongside vector search, self-hosted from one binary or on its managed cloud. Chosen by 18 makers' products we track. - [Weaviate](https://howmakersbuild.com/tools/weaviate) (Open-source vector store): Hybrid search that mixes keyword and vector results, with built-in modules that can create embeddings for you. Chosen by 16 makers' products we track. - [turbopuffer](https://howmakersbuild.com/tools/turbopuffer) (Managed vector store): Very large or many-tenant indexes where storing everything on object storage keeps cost down. Chosen by 8 makers' products we track. - [LLM Observability & Evals](https://howmakersbuild.com/tasks/llm-observability): Open-source platform you can self-host, or a hosted eval product. Open-source tools keep traces on your own servers and cost nothing to run yourself, but you maintain them. Hosted eval platforms give you polished experiment and scoring workflows for a monthly bill, with your data in their cloud. - [Langfuse](https://howmakersbuild.com/tools/langfuse) (Open-source platform): Tracing, prompt management and evals in one tool you can run on your own server for free. Chosen by 33 makers' products we track. - [Arize Phoenix](https://howmakersbuild.com/tools/arize-phoenix) (Open-source platform): OpenTelemetry-based tracing and evals you can start locally in a notebook, then self-host or move to its cloud. - [Helicone](https://howmakersbuild.com/tools/helicone) (Proxy logging): Getting request logs, costs and latency by changing one base URL, with no SDK instrumentation. Chosen by 13 makers' products we track. - [LangSmith](https://howmakersbuild.com/tools/langsmith) (Hosted eval platform): Apps built on LangChain or LangGraph, where tracing works with almost no setup. Chosen by 11 makers' products we track. - [Braintrust](https://howmakersbuild.com/tools/braintrust) (Hosted eval platform): Eval-driven work — scoring outputs and comparing prompts and models side by side in experiments. Chosen by 5 makers' products we track. ## Shipping - [Hosting](https://howmakersbuild.com/tasks/hosting): What shape is your app? A frontend or serverless app fits a frontend cloud. Long-running servers, background workers and databases fit a PaaS or container platform. A hyperscaler cloud has every building block but leaves you to assemble them; a platform or framework that deploys into your own cloud account sits between, with platform-style deploys and the resources on your own cloud bill. - [Vercel](https://howmakersbuild.com/tools/vercel) (Frontend cloud): Next.js apps — zero-config deploys and preview URLs. - [Netlify](https://howmakersbuild.com/tools/netlify) (Frontend cloud): Static sites and JAMstack apps on any framework. Chosen by 241 makers' products we track. - [Cloudflare](https://howmakersbuild.com/tools/cloudflare) (Edge platform): Cheap global hosting with Workers, plus database (D1) and storage (R2) on the same platform. Chosen by 375 makers' products we track. - [Railway](https://howmakersbuild.com/tools/railway) (PaaS): Servers, workers and databases deployed together, with almost no configuration. Chosen by 112 makers' products we track. - [Render](https://howmakersbuild.com/tools/render) (PaaS): A Heroku-style platform for web services, cron jobs and Postgres. Chosen by 128 makers' products we track. - [Fly.io](https://howmakersbuild.com/tools/fly-io) (Container platform): Docker apps that should run close to users in many regions. Chosen by 66 makers' products we track. - [Amazon Web Services](https://howmakersbuild.com/tools/aws) (Hyperscaler cloud): The widest catalog of building blocks — Lambda, containers, managed databases, queues — to assemble your own infrastructure. Chosen by 386 makers' products we track. - [Hetzner Cloud](https://howmakersbuild.com/tools/hetzner) (Your own server): A lot of CPU and RAM per euro, in European, US and Singapore regions. Chosen by 100 makers' products we track. - [Coolify](https://howmakersbuild.com/tools/coolify) (Your own server): A Heroku- or Vercel-style dashboard on your own server, with many one-click services. Chosen by 23 makers' products we track. - [Image & Video Hosting](https://howmakersbuild.com/tasks/media-hosting): Image pipeline, video streaming, or a raw CDN bundle. Image services resize, crop and convert through the URL. Video needs encoding and adaptive streaming, which is a separate job. A CDN bundle gives storage, delivery and basic optimization at low per-GB cost, with less done for you. Or process images yourself, with a library in your code or an image server you host. - [Cloudinary](https://howmakersbuild.com/tools/cloudinary) (Image pipeline): Apps with lots of user-uploaded images and video that need crops, overlays and format conversion through URLs. Chosen by 47 makers' products we track. - [ImageKit](https://howmakersbuild.com/tools/imagekit) (Image pipeline): Adding resizing and optimization in front of storage you already have, such as an S3 bucket. Chosen by 20 makers' products we track. - [Mux](https://howmakersbuild.com/tools/mux) (Video streaming): Apps where video is the product — uploads encoded and streamed adaptively with a player and playback analytics. Chosen by 23 makers' products we track. - [Bunny.net](https://howmakersbuild.com/tools/bunny) (CDN bundle): Low per-GB cost for storage, CDN delivery, image optimization and video streaming from one vendor. - [sharp](https://howmakersbuild.com/tools/sharp) (Process it yourself): Resizing, cropping and converting images to WebP or AVIF inside your own Node code, at upload or build time. - [imgproxy](https://howmakersbuild.com/tools/imgproxy) (Process it yourself): Cloudinary-style resizing through URLs on a server you run, placed in front of your bucket and behind a CDN. - [Cloudflare](https://howmakersbuild.com/tools/cloudflare) (CDN bundle): Image resizing and video streaming from the same account that already runs your DNS and CDN, through Cloudflare Images and Stream. - [Error Monitoring](https://howmakersbuild.com/tasks/error-monitoring): Hosted service or self-hosted tracker. Hosted services give you setup in minutes and pricing by event volume; self-hosting keeps error data on your servers for the cost of running and upgrading it. - [Sentry](https://howmakersbuild.com/tools/sentry) (Hosted service): Broad SDK coverage across languages and frameworks, with performance tracing next to errors. Chosen by 162 makers' products we track. - [Rollbar](https://howmakersbuild.com/tools/rollbar) (Hosted service): Error grouping tied to releases, with session replays linked to errors. Chosen by 6 makers' products we track. - [GlitchTip](https://howmakersbuild.com/tools/glitchtip) (Self-hosted tracker): Self-hosting error tracking on a small server while using Sentry's SDKs. ## Growth - [Web Analytics](https://howmakersbuild.com/tasks/web-analytics): Cookie-free dashboard or full marketing suite. Cookie-free tools show pageviews, referrers, UTM campaigns and simple goals on one page and usually need no consent banner. Google Analytics is free and goes much deeper into campaigns and ads, but sets cookies, so EU visitors need a consent banner and some of them are never counted. - [Plausible Analytics](https://howmakersbuild.com/tools/plausible) (Cookie-free dashboard): One page of traffic numbers with no cookie banner, with funnels and revenue goals on higher plans. Chosen by 198 makers' products we track. - [Umami](https://howmakersbuild.com/tools/umami) (Cookie-free dashboard): Open-source website analytics you can self-host free, with a cloud free plan up to 100,000 events a month. Chosen by 48 makers' products we track. - [Fathom Analytics](https://howmakersbuild.com/tools/fathom) (Cookie-free dashboard): A cookie-free dashboard with up to 50 sites on every plan. Chosen by 40 makers' products we track. - [Cloudflare Web Analytics](https://howmakersbuild.com/tools/cloudflare-web-analytics) (Built into your host): Free cookie-free visit counts for sites on any host, set up in minutes. Chosen by 92 makers' products we track. - [Google Analytics](https://howmakersbuild.com/tools/google-analytics) (Marketing suite): Free, deep campaign and audience reporting tied to Google Ads and Search Console. Chosen by 410 makers' products we track. - [Product Analytics](https://howmakersbuild.com/tasks/product-analytics): All-in-one platform or dedicated analytics tool. All-in-one platforms put event analytics next to session replay, feature flags and experiments, billed per product. Dedicated tools go deeper on funnels, retention and cohorts, with replay and flags added around them. Every option needs events wired into your code. - [PostHog](https://howmakersbuild.com/tools/posthog) (All-in-one platform): One tool for analytics, session replay, feature flags, experiments and surveys, each with its own free monthly allowance. Chosen by 321 makers' products we track. - [Mixpanel](https://howmakersbuild.com/tools/mixpanel) (Dedicated analytics): Mature funnel, retention and cohort analysis on event data. Chosen by 108 makers' products we track. - [Amplitude](https://howmakersbuild.com/tools/amplitude) (Dedicated analytics): Funnel, retention and cohort analysis, with session replay and experiments on every plan, including free. Chosen by 87 makers' products we track. - [Newsletter](https://howmakersbuild.com/tasks/newsletter): Growth platform or writer's tool. beehiiv and Kit bring growth and monetization features — referrals, ads, automations. Buttondown and Substack keep things simple for writing and sending; Substack adds a built-in reader network but takes a cut of paid subscriptions. - [beehiiv](https://howmakersbuild.com/tools/beehiiv) (Growth platform): A newsletter you want to grow and monetize, with referrals, a recommendation network and ads. Chosen by 14 makers' products we track. - [Kit](https://howmakersbuild.com/tools/kit) (Growth platform): Creators who need automations and tagged sequences, like a course launch funnel. Chosen by 67 makers' products we track. - [Buttondown](https://howmakersbuild.com/tools/buttondown) (Writer's tool): Developers who write in Markdown and want a small, privacy-friendly tool with an API. - [Substack](https://howmakersbuild.com/tools/substack) (Writer's tool): Writers who want discovery from Substack's own reader network and free sending. Chosen by 16 makers' products we track. - [Lifecycle Email](https://howmakersbuild.com/tasks/lifecycle-email): Email-only or multichannel. Email-focused tools keep sequences simple and cheap to start; messaging automation platforms add push, SMS and in-app messages driven by the same behavior data, with more setup and a higher bill. - [Loops](https://howmakersbuild.com/tools/loops) (Email for SaaS): Event-triggered onboarding sequences and transactional email from one tool, with a free tier to start. Chosen by 219 makers' products we track. - [Userlist](https://howmakersbuild.com/tools/userlist) (Email for SaaS): B2B SaaS where messages depend on both the user and the company account they belong to. Chosen by 8 makers' products we track. - [Customer.io](https://howmakersbuild.com/tools/customer-io) (Multichannel automation): Complex behavior-driven journeys across email, push, SMS and in-app messages. Chosen by 48 makers' products we track. ## Building with AI - [AI Coding Assistants](https://howmakersbuild.com/tasks/ai-coding-assistant): Where the agent works. AI-first editors put completions, chat and agents inside the editor where you watch every change; terminal agents take a task and work across the repo, running commands, with you reviewing the result; editor plugins add AI to the editor you already use without switching; agent managers run several agents in parallel, each on its own copy of the repo, and help you review and merge their work. - [Cursor](https://howmakersbuild.com/tools/cursor) (AI-first editor): A VS Code-style editor with AI completions, chat and agents built in, keeping your VS Code extensions. Chosen by 538 makers' products we track. - [Devin Desktop (formerly Windsurf)](https://howmakersbuild.com/tools/windsurf) (AI-first editor): A VS Code-style editor with unlimited Tab completions, an in-editor agent, and one place to manage local and cloud Devin agents. Chosen by 43 makers' products we track. - [Claude Code](https://howmakersbuild.com/tools/claude-code) (Terminal agent): Handing off multi-step tasks — features, refactors, bug hunts — to an agent that reads the repo, edits files and runs commands. Chosen by 738 makers' products we track. - [OpenAI Codex](https://howmakersbuild.com/tools/openai-codex) (Terminal agent): Delegating tasks to an agent in the terminal, IDE or desktop app, or to cloud tasks that run in the background and open pull requests. Chosen by 136 makers' products we track. - [OpenCode](https://howmakersbuild.com/tools/opencode) (Terminal agent): An open-source coding agent that works with almost any model provider, local models, or an existing Copilot or ChatGPT subscription. Chosen by 42 makers' products we track. - [GitHub Copilot](https://howmakersbuild.com/tools/github-copilot) (Editor plugin): Adding AI completions, chat and agents to the editor you already use, including VS Code and JetBrains IDEs, plus pull request reviews on GitHub. Chosen by 26 makers' products we track. - [AI App Builders](https://howmakersbuild.com/tasks/ai-app-builder): Whether you'll keep the code. Hosted builders take you from prompt to deployed app with a backend wired up, but you work inside their environment and pay by credits; code-first generators give you components to pull into your own repo; a cloud IDE with an agent sits in between and lets you keep editing the code by hand. - [Lovable](https://howmakersbuild.com/tools/lovable) (Hosted app builder): Full-stack web apps from chat, with auth and database through a built-in Supabase connection. Chosen by 73 makers' products we track. - [Base44](https://howmakersbuild.com/tools/base44) (Hosted app builder): Full-stack web apps and internal tools from chat, with database, auth and hosting built into the platform, so there is no outside backend account to set up. Chosen by 11 makers' products we track. - [Bolt](https://howmakersbuild.com/tools/bolt-new) (Hosted app builder): Generating and running a full-stack app entirely in the browser, with the code visible and editable as it is written. Chosen by 24 makers' products we track. - [v0](https://howmakersbuild.com/tools/v0) (Code-first generator): React and Next.js UI and apps that you pull into your own repository or deploy on Vercel. Chosen by 41 makers' products we track. - [Replit](https://howmakersbuild.com/tools/replit) (Cloud IDE with agent): Building with an AI agent in a full cloud IDE, in many languages, with hosting included. Chosen by 34 makers' products we track. ## Pricing at your scale - [AI Coding Assistants pricing](https://howmakersbuild.com/pricing/ai-coding-assistant): what each AI coding assistants tool costs per month at a usage you set — developers — cheapest plan first, from each vendor's pricing page. - [Authentication pricing](https://howmakersbuild.com/pricing/authentication): what each authentication tool costs per month at a usage you set — monthly active users — cheapest plan first, from each vendor's pricing page. - [Background Jobs & Cron pricing](https://howmakersbuild.com/pricing/background-jobs): what each background jobs & cron tool costs per month at a usage you set — job runs or steps per month, job compute time per month (small machine, 0.5 vcpu) — cheapest plan first, from each vendor's pricing page. - [Database pricing](https://howmakersbuild.com/pricing/database): what each database tool costs per month at a usage you set — database size, compute hours per month (a small instance, always on ≈ 730), data transfer per month, document or row reads per month (for per-operation pricing), document or row writes per month (for per-operation pricing) — cheapest plan first, from each vendor's pricing page. - [Error Monitoring pricing](https://howmakersbuild.com/pricing/error-monitoring): what each error monitoring tool costs per month at a usage you set — errors per month — cheapest plan first, from each vendor's pricing page. - [File Storage pricing](https://howmakersbuild.com/pricing/file-storage): what each file storage tool costs per month at a usage you set — stored, downloaded per month — cheapest plan first, from each vendor's pricing page. - [Hosting pricing](https://howmakersbuild.com/pricing/hosting): what each hosting tool costs per month at a usage you set — requests per month, data transfer out per month, serverless function compute (gb-hours; about 70 ms at 1 gb per request), always-on app servers (1 vcpu, 2 gb ram each), developers on the team — cheapest plan first, from each vendor's pricing page. - [Image & Video Hosting pricing](https://howmakersbuild.com/pricing/media-hosting): what each image & video hosting tool costs per month at a usage you set — images stored (originals, about 500 kb each), new image variants generated per month (sizes and formats not yet cached), images delivered per month (about 100 kb each optimized) — cheapest plan first, from each vendor's pricing page. - [Lifecycle Email pricing](https://howmakersbuild.com/pricing/lifecycle-email): what each lifecycle email tool costs per month at a usage you set — contacts, emails per month — cheapest plan first, from each vendor's pricing page. - [LLM API pricing](https://howmakersbuild.com/pricing/llm-api): what each LLM API tool costs per month at a usage you set — input tokens per month, output tokens per month — cheapest plan first, from each vendor's pricing page. - [LLM Observability & Evals pricing](https://howmakersbuild.com/pricing/llm-observability): what each LLM observability & evals tool costs per month at a usage you set — traces per month — cheapest plan first, from each vendor's pricing page. - [Newsletter pricing](https://howmakersbuild.com/pricing/newsletter): what each newsletter tool costs per month at a usage you set — subscribers, paid subscription revenue per month (platform's cut only, card fees excluded) — cheapest plan first, from each vendor's pricing page. - [Payments pricing](https://howmakersbuild.com/pricing/payments): what each payments tool costs per month at a usage you set — revenue per month, payments per month — cheapest plan first, from each vendor's pricing page. - [Product Analytics pricing](https://howmakersbuild.com/pricing/product-analytics): what each product analytics tool costs per month at a usage you set — events per month — cheapest plan first, from each vendor's pricing page. - [Transactional Email pricing](https://howmakersbuild.com/pricing/transactional-email): what each transactional email tool costs per month at a usage you set — emails per month — cheapest plan first, from each vendor's pricing page. - [Vector Database pricing](https://howmakersbuild.com/pricing/vector-database): what each vector database tool costs per month at a usage you set — vectors stored (1,536 dimensions, about 6 gb per million), queries per month, vectors written or updated per month — cheapest plan first, from each vendor's pricing page. - [Web Analytics pricing](https://howmakersbuild.com/pricing/web-analytics): what each web analytics tool costs per month at a usage you set — pageviews per month — cheapest plan first, from each vendor's pricing page. ## Optional - Head-to-head pages: /compare/-vs- for the pairs makers search, e.g. [Firebase vs Supabase](https://howmakersbuild.com/compare/firebase-vs-supabase). - Per framework or language: /tasks//for/, e.g. [Authentication for Next.js](https://howmakersbuild.com/tasks/authentication/for/nextjs). - [All tools](https://howmakersbuild.com/tools) - [Products we track](https://howmakersbuild.com/products)