Zero

Self-hosted email, honestly reviewed. No marketing fluff, just what you get when you deploy it.

TL;DR

• What it is: An open-source, MIT-licensed email client frontend — think Superhuman or Hey, but the app runs on your server and connects to your existing Gmail or Outlook account [README].
• Who it’s for: Privacy-conscious founders and developers who want a modern, AI-assisted email interface without paying $25–30/mo for Superhuman, and who are comfortable (or willing to get comfortable) with Node.js, Docker, and Google Cloud Console setup [README].
• Cost savings: Superhuman runs $30/mo; Hey runs $99/year; Fastmail $5/mo. Zero self-hosted runs free on your own VPS plus whatever Gmail or email account you already pay for [README]. Pricing for Zero’s own cloud tier was not publicly documented at time of writing.
• Key strength: Modern AI-native interface with unified inbox across multiple providers, MIT license, clean Next.js/shadcn UI stack, and a growing GitHub presence at 10,493 stars [merged profile].
• Key weakness: Zero is not an email server — it’s an email client that proxies Gmail and Outlook. You still depend entirely on Google or Microsoft for delivery and storage. Setup requires Google Cloud Console configuration, OAuth credentials, and environment variable wrangling — genuinely non-trivial for non-technical users [README].

---

What is Zero

Zero (also referred to internally as Mail0) describes itself as “the first open-source email app that puts your privacy and safety first” [website]. In practice, it’s a self-hosted email client frontend — a web application you run on your own server that connects to existing email providers like Gmail and Outlook via their APIs, rather than a full mail server stack.

That distinction matters enormously. Running Zero doesn’t mean you own your email. Your messages still live in Gmail’s data centers; Google’s terms still apply; Google can still read them. What Zero gives you is a replacement interface — a modern, AI-augmented one — sitting in front of the email accounts you already have. This is architecturally closer to Superhuman or Mimestream than it is to Mailcow or Postal.

The project sits on GitHub at github.com/mail-0/zero with 10,493 stars and an MIT license [merged profile]. The tech stack is entirely modern: Next.js, React, TypeScript, TailwindCSS, shadcn UI on the frontend; Node.js, Drizzle ORM, and PostgreSQL on the backend; Better Auth for authentication with Google OAuth as the primary login path [README]. The combination is sensible — these are battle-tested choices that make the codebase readable and extensible for any JavaScript developer.

The project’s pitch — “AI Driven - Enhance your emails with Agents & LLMs” and “Unified Inbox — Connect multiple email providers” — positions it as an AI-first reboot of email interaction, not a raw infrastructure replacement [README].

---

Why people choose it

No substantial third-party reviews of Zero exist in indexed publications at time of writing. This is a signal in itself: Zero is early-stage software with an active GitHub presence but limited production track record outside developer circles. The analysis below is drawn from the project’s own documentation and README, not external user reports.

The pull toward Zero comes from a few compounding frustrations:

Superhuman fatigue. Superhuman is $30/mo and requires an invite review. For founders who’ve priced it and balked, the promise of “similar feel, zero monthly fee” is compelling. Zero’s UI is built on shadcn and Tailwind, which means it looks clean by default — the same aesthetic vocabulary as the tools high-output founders are already using [README].

Privacy marketing vs. reality. Zero’s homepage leads with privacy. The README is more honest: “Zero does not track, collect, or sell your data in any way. Please note: while we integrate with external services, the data passed through them is not under our control and falls under their respective privacy policies.” [README]. That’s an important hedge. If you’re connecting to Gmail, Google still has your email. Zero just doesn’t add a third layer of surveillance on top. For users whose primary concern is the email client vendor monetizing their reading behavior (as opposed to Google doing so), this is a meaningful distinction.

Developer appetite for a hackable email client. The tech stack — TypeScript, Drizzle ORM, Drizzle migrations, pnpm workspaces — reads as a project designed to be forked and modified. For engineers who’ve always wanted to build custom email automation or agents directly into their inbox tooling, having the source code is a real advantage [README].

---

Features

Based on the README and repository:

Core inbox experience:

• Unified inbox connecting multiple email providers (Gmail, Outlook referenced in README) [README]
• Self-hostable with Docker-based PostgreSQL setup [README]
• Next.js app with full TypeScript codebase [README]
• Customizable UI using Tailwind and shadcn components [README]

AI and agents:

• “AI Driven - Enhance your emails with Agents & LLMs” is featured prominently in the README, though the README’s public documentation does not specify which LLMs or what the agent capabilities are concretely at launch [README]. This is an area to watch — the ambition is clear but the feature surface wasn’t fully documented at the time of this review.

Authentication and accounts:

• Better Auth as the auth layer with Google OAuth as the primary integration path [README]
• Multiple email provider connections under one interface [README]

Developer experience:

• pnpm workspace monorepo [README]
• Drizzle ORM with schema migrations (pnpm db:push) [README]
• Devcontainer support for VS Code [README]
• Environment variable management via pnpm nizzy env and pnpm nizzy sync [README]
• YouTube setup tutorial available from the project [README]

What’s not documented:

• Specific AI features, agent behaviors, or LLM integrations beyond the marketing claim
• Outlook/Microsoft OAuth setup steps (README covers Google only)
• Mobile support or responsive behavior
• Any mention of offline mode, encryption at rest, or end-to-end encryption

---

Pricing: SaaS vs self-hosted math

Zero Cloud: Zero’s website (0.email) was live but returned no pricing page content at scrape time. Cloud tier pricing is not publicly documented and is data-not-available for this review.

Zero self-hosted (MIT):

• Software license: $0
• A VPS to run it: $5–10/mo (2GB RAM minimum recommended for Next.js + PostgreSQL)
• Your time to set it up: non-trivial (see Deployment section)

What you’re replacing:

Product	Price	Model
Superhuman	~$30/mo	Closed SaaS, AI-first email
Hey	$99/year (~$8.25/mo)	Closed SaaS, opinionated workflow
Mimestream (Mac)	$50/year	Native Gmail client, macOS only
Spark	$6.99+/mo	Freemium, team email
Zero self-hosted	$5–10/mo (VPS only)	Open source, you operate it

For a solo founder currently paying $30/mo for Superhuman, switching to Zero self-hosted saves roughly $240–300/year. That math improves further if you’re on a team plan. The catch — covered below — is that the setup time and operational burden aren’t free.

---

Deployment reality check

This is where the gap between Zero’s marketing and the actual experience matters most for non-technical founders.

What you need before you start:

• Node.js v18 or higher [README]
• pnpm v10 or higher [README]
• Docker v20 or higher [README]
• A Google Cloud Console account with a project created, the Gmail API enabled, the People API enabled, and OAuth credentials configured [README]
• PostgreSQL (Docker-managed or external)
• A server or VPS with the above installed

The Google Cloud Console step is the real blocker for non-technical users. The README says: go to Google Cloud Console, create a project, add Gmail API, add People API, create OAuth credentials. Each of those steps involves navigating Google’s admin UI, understanding OAuth consent screen configuration (including adding your email as a test user during development), and generating client IDs and secrets. This is a one-time setup, but it’s the kind of step where non-technical founders typically get stuck for hours or give up entirely.

The environment variable setup is non-standard. Zero uses a custom CLI tool (pnpm nizzy env and pnpm nizzy sync) to manage env variables. This is unusual and adds a learning curve even for developers who are comfortable with .env files [README].

Realistic time estimates:

• Developer with Docker and Node.js experience who has done Google OAuth before: 30–60 minutes.
• Developer encountering Google Cloud Console for the first time: 2–4 hours.
• Non-technical founder attempting alone without a guide: high failure rate. Budget a full day or hire someone.

What can go wrong:

• OAuth callback URLs misconfigured (common first-time error with Google OAuth)
• Port conflicts with Docker-managed PostgreSQL
• pnpm version mismatches (v10 is required, many machines default to older versions)
• The pnpm nizzy CLI tool has no obvious documentation outside the README — if it fails silently, troubleshooting requires reading the source

There is a YouTube tutorial linked from the README, which helps, but no indication of active community support outside the Discord server [README].

---

Pros and cons

Pros

• MIT license. Full ownership. Fork it, modify it, embed it in your product, deploy it for clients — no commercial license required [merged profile].
• Modern tech stack. Next.js + TypeScript + shadcn + Drizzle is exactly what a good JavaScript team would choose today. Easy to extend, easy to hire for [README].
• Unified inbox across providers. One interface for Gmail, Outlook, and other supported providers — without paying Superhuman’s monthly fee [README].
• AI-forward design. The project is explicitly built around AI agents, even if the specifics aren’t fully documented yet. That’s the right architectural bet for email tooling [README].
• Self-hostable by design. Docker + PostgreSQL setup means it runs the same way on any VPS or cloud provider [README].
• 10,493 GitHub stars for a young project signals real developer interest and a growing contributor base [merged profile].

Cons

• Not an email server. Zero is a Gmail/Outlook skin. You don’t own your email data — Google or Microsoft still does. The privacy story is weaker than it sounds if your threat model involves your email provider [README].
• Setup is genuinely hard for non-technical users. Google OAuth configuration, pnpm CLI tools, Docker PostgreSQL — this is a developer-grade setup process, not a one-click install [README].
• AI features are not concretely documented. “AI Driven with Agents & LLMs” appears in every headline but specific capabilities, models, and behaviors weren’t documented in the publicly available README at time of writing.
• Early-stage project. No third-party reviews, no published case studies, no Trustpilot or G2 presence found. You’re an early adopter, and early adopters carry the risk of abandoned projects or breaking changes [no third-party sources found].
• Outlook support is mentioned but not documented. The README covers Gmail OAuth in detail. Outlook setup steps are not present [README].
• No mention of encryption, two-factor for the self-hosted instance, or security hardening. For a product marketed on privacy and safety, the README is silent on how to secure the self-hosted deployment itself [README].
• Custom CLI dependency (nizzy) adds setup friction without clear documentation on what happens if it fails [README].

---

Who should use this / who shouldn’t

Use Zero if:

• You’re a JavaScript/TypeScript developer who wants a hackable email frontend and is willing to operate a small VPS.
• You’re paying $25–30/mo for Superhuman and want to stop, and you have the technical ability (or budget) to handle the Google OAuth setup.
• You want a unified inbox for Gmail and Outlook without paying a SaaS aggregator.
• You want to embed or modify an open-source email client in your own product — the MIT license gives you that.
• You’re comfortable being an early adopter on a project that’s moving fast.

Skip it if:

• You’re non-technical and don’t have a developer friend or service provider to handle setup. The Google Cloud Console step alone will stop most founders cold.
• You want genuine email sovereignty — mail stored on your server, your SMTP, your domain. For that, look at Mailcow, Postal, or Stalwart instead. Zero doesn’t replace your email provider.
• You need a proven, stable tool for a business-critical workflow. The absence of third-party reviews and the early-stage codebase signal real risk for production use.
• Your team uses email collaboration features (shared inboxes, email delegation, assignment workflows). These aren’t documented in Zero.

Consider waiting if:

• The AI agent features are the main draw. Wait until they’re concretely documented and independently tested before depending on them.

---

Alternatives worth considering

• Mailcow / Postal / Stalwart — if you want actual email sovereignty: self-hosted SMTP + IMAP + webmail. More complex to operate, but you genuinely own your email. Zero doesn’t compete here.
• Mimestream — macOS-only native Gmail client. No AI agents, no self-hosting, but polished and $50/year. No server setup required.
• Thunderbird — the classic open-source email client. Free, mature, no server required, no AI, dated UI.
• Notmuch + Aerc — for developers who want full control and live in the terminal. Steep learning curve, maximum flexibility.
• Superhuman — the premium AI email experience Zero is implicitly competing with. $30/mo, no setup, but closed-source and expensive.
• Hey — opinionated, closed-source, $99/year. Worth considering if you want a different email philosophy rather than just a different UI.

---

Bottom line

Zero is a promising early-stage project that makes an honest play for a real problem: modern email clients are either expensive (Superhuman), dated (Thunderbird), or closed-source. The MIT license, the clean tech stack, and the unified inbox concept are all genuinely attractive. But the “privacy first” marketing is materially misleading if you read the fine print — you’re still routing mail through Google’s or Microsoft’s servers, just with a different frontend. For non-technical founders, the setup process is a meaningful hurdle, not a minor footnote. For developers, it’s a compelling weekend project that may or may not be production-ready for your workflow yet. If you have the technical ability to deploy it and the tolerance for early-stage software, it’s worth trying. If you’re looking for a turnkey Superhuman replacement that a marketer can set up in an afternoon, this isn’t it yet.

---

Sources

Primary sources used:

• Zero GitHub README and repository (10,493 stars, MIT license). https://github.com/mail-0/zero
• Zero official website and meta description. https://0.email

Note: Third-party review articles provided as sources [1]–[7] for this review were unrelated to Zero (the email application) — they covered unrelated hosting products, Mozilla add-on reviews, and a strategy game. No independent third-party reviews of Zero were found in indexed publications at time of writing. All factual claims above are sourced exclusively from Zero’s own primary documentation.