What Superhuman actually does
Superhuman is an AI-powered premium email client founded in 2014–2015 by Rahul Vohra, Conrad Irwin, and Vivek Sodera. It was acquired by Grammarly on July 1, 2025 (TechCrunch) for an undisclosed price — analyst estimates range from $300M to $500M. Grammarly's parent company subsequently rebranded to 'Superhuman' in October 2025. Pre-acquisition valuation was $825M from a Series C in August 2021. At the time of acquisition, Superhuman had approximately 70,000 paying users and ~$35–36.5M ARR (Sacra June 2025 estimate).
Superhuman's core value proposition is speed — the client is engineered to make every email interaction faster through keyboard shortcuts, instant search, a command palette (Cmd+K for everything), and AI that drafts responses in your writing style. The mandatory onboarding call (a paid white-glove setup with a Superhuman team member) was a signature early feature that drove viral word-of-mouth but is now criticized as a sales process rather than genuine onboarding.
The product's fundamental limitations are significant: Gmail and Outlook OAuth only — no Fastmail, iCloud, or generic IMAP support. No shared inbox. No email delegation. At $30/user/mo for these constraints, the value proposition is increasingly questioned, particularly post-Grammarly acquisition where Superhuman's roadmap is now driven by Grammarly's broader AI writing strategy rather than pure email optimization.
Keyboard-first email interface
The entire application is designed around keyboard shortcuts — navigation, archiving, labeling, composing, and searching all have dedicated key bindings. The command palette (Cmd+K) provides instant access to any action. This is Superhuman's primary UX differentiator and the hardest feature to replicate authentically.
Split inbox with auto-triage
Superhuman automatically separates incoming email into Important (1-on-1 and team emails) and Other (newsletters, notifications) using Gmail's categorization. The split inbox eliminates the need to manually process low-priority email.
AI drafting and response suggestions
AI generates reply drafts based on email context and the user's writing history. 'Ask AI' generates full email drafts from a short prompt. Auto Drafts (Business tier) proactively drafts replies before you open emails. These features are the primary reason teams evaluate Superhuman in 2025.
Gmail and Outlook OAuth integration
Superhuman connects via OAuth to Gmail (Google Workspace and personal) and Microsoft Outlook/Exchange. The entire product depends on these two email provider integrations — no other providers are supported.
Snooze, follow-up reminders, and read status tracking
Snooze emails to reappear at a specific time. Remind me if no reply within N days. See who opened your email and when via read receipt tracking. These features are the practical productivity drivers beyond the keyboard-shortcut experience.
Cross-platform sync
Native macOS app, iOS app, and web app with sub-second load times. All state synced in real-time across devices. The 'instant loading' experience is a core product promise — achieved via local SQLite/LMDB caching of the email index.
Superhumanpricing & limits
Based on 10 users on Starter plan at $300/year each
Where Superhuman falls short
Most expensive email client on the market at $30/user/mo
Independent 2026 comparisons (Spark vs. Superhuman) identify Superhuman as the most expensive email client available — at 3–4x the price of alternatives like Spark ($4.99/mo), Hey ($99/yr), or Mimestream (free/one-time). A 10-person team pays $3,000/yr for keyboard shortcuts and AI email features. For teams that don't email-intensively, this cost is impossible to justify.
Gmail and Outlook lock-in — no IMAP or alternative providers
Superhuman works exclusively with Gmail (personal and Google Workspace) and Microsoft Outlook/Exchange. Teams using Fastmail, iCloud, ProtonMail, Tutanota, or self-hosted email servers cannot use Superhuman at all. The alfred_ 2025 review documents this as a critical adoption blocker for privacy-conscious teams and organizations with custom email infrastructure.
No shared inbox or delegation even at $30/user
Support teams, customer success teams, and executive assistants that need shared inbox functionality (multiple people managing a single email address like support@company.com) cannot use Superhuman. The alfred_ 2025 review specifically calls out paying '$30/user × 5 with no shared inbox or delegation' as the primary frustration. Support use cases are simply not served by Superhuman's individual-focused architecture.
Mandatory onboarding feels like a sales process
Superhuman's original white-glove onboarding call was a signature product decision that drove viral word-of-mouth. Reddit threads from 2024 document growing frustration with what users describe as a 30-minute sales pitch disguised as onboarding. The mandatory setup call creates friction for team rollouts where dozens of users must each complete their own onboarding session.
Roadmap uncertainty post-Grammarly acquisition
Grammarly's acquisition of Superhuman in July 2025 and subsequent parent company rebrand to 'Superhuman' in October 2025 signals a strategy shift. Grammarly's core business is AI writing assistance — integrating Superhuman into a broader AI writing suite may deprioritize features that email-first power users value. Teams evaluating Superhuman must consider a roadmap that may be driven by Grammarly's writing tool strategy rather than pure email optimization.
Key features to replicate
The core feature set any Superhuman alternative needs — plus what you can improve on.
Keyboard-first interface with command palette
A comprehensive keyboard shortcut system where every action (archive, reply, label, snooze, search, navigate) has a key binding. The command palette (Cmd+K) provides instant action search — type 'snooze' and select a time without reaching for the mouse. This is primarily a frontend engineering challenge: a well-designed shortcut system with conflict resolution, customizable bindings, and discoverability (hint overlays on 'J/K' for navigate, '?' for help).
Gmail and Outlook OAuth integration with local cache
OAuth 2.0 authentication with Google and Microsoft. The Gmail API (REST) and Microsoft Graph API (REST) provide read/write access to messages, labels, threads, and contacts. To achieve 'instant' loading, synchronize the email inbox to a local SQLite or LMDB cache on startup, then serve the UI from the cache while syncing in the background. This local-first architecture is the engineering core of Superhuman's speed.
Split inbox and auto-triage
Gmail's API provides message categories (Primary, Social, Promotions, Updates, Forums). Use these categories plus sender history analysis to classify emails as Important vs. Other. Store the user's triage decisions in a local SQLite database to improve classification over time. The split inbox is a rendering concern — show two lists from the same email data with different filter criteria applied.
AI drafting with writing style learning
Fine-tuning a writing style model requires sample emails from the user's Sent folder. In practice, few-shot prompting with GPT-4o or Claude — providing 10–20 recent sent emails as context examples — produces good style matching without model training. Draft generation is triggered by pressing a shortcut key in compose mode. The AI call is asynchronous and streams the draft token-by-token for perceived speed.
Snooze and follow-up reminder system
Snooze moves an email from the inbox temporarily, returning it at a specified time. Implement by adding a label to the email and scheduling a job (Inngest or BullMQ) to remove the label at the snooze time. Follow-up reminders check the thread for replies — if no reply arrives by the reminder time, the job re-surfaces the email. Both require access to the email thread via the Gmail/Graph API.
Read receipt tracking
Embed a 1x1 pixel tracking image in outgoing emails hosted on your server. When the recipient opens the email, their mail client loads the tracking pixel, logging the request with email ID, timestamp, and approximate location via IP geolocation. Note: this is blocked by many email privacy features (Apple Mail Privacy Protection, Gmail image proxy). Implement with accurate expectations about coverage (~40–60% of emails tracked).
Cross-platform sync with instant load
A background sync service polls the Gmail and Graph APIs for new emails and pushes changes to a PostgreSQL backend. The macOS and iOS apps maintain a local SQLite cache that hydrates the UI instantly on launch. The web app uses a combination of server-side rendering for initial load and WebSocket for real-time inbox updates. The 'instant' experience requires the local cache — without it, every inbox load requires an API call.
Technical architecture
A Superhuman alternative is a local-first email client with a cloud sync backend. The core technical challenge is achieving the 'feel of speed' — sub-100ms keyboard interactions and instant inbox loading — which requires local SQLite caching synchronized with the Gmail/Outlook APIs. The AI layer is relatively straightforward (GPT-4o or Claude with few-shot prompting). The platform-specific work (native macOS app, iOS app) is significant and required for Superhuman's primary value proposition.
Web frontend
Next.js App Router, React + Vite, Electron (web wrapper)
Recommended: React + Vite for the web client — email is highly interactive (keyboard shortcuts, real-time updates) and benefits from a fast client-side rendering approach. Next.js adds SSR complexity that doesn't benefit an email client UX.
Desktop app
Electron + React, Tauri + WebView, native Swift (macOS)
Recommended: Electron with the web React app wrapped — enables code sharing between web and desktop. Tauri (Rust) provides better performance but requires separate React and Rust codebases. Native Swift is best UX but doubles the development work.
Email provider integration
Gmail API (Google), Microsoft Graph API (Microsoft), Nylas (unified API)
Recommended: Build native Gmail and Microsoft Graph integrations — covers 95%+ of professional email users. Nylas provides a unified API that also supports IMAP, enabling support for Fastmail, iCloud, and other providers. Nylas costs ~$0.004/message but saves months of integration work.
Local cache
SQLite (local), LMDB, IndexedDB (web)
Recommended: SQLite via the better-sqlite3 npm package for desktop; IndexedDB (via Dexie.js) for web. Cache the last 3,000–5,000 emails with full metadata and body. This is the 'instant load' foundation.
Backend API
NestJS, Next.js Route Handlers, Express
Recommended: NestJS — handles the email sync service, AI generation endpoints, snooze/reminder jobs, and account management with clean module separation.
AI integration
OpenAI GPT-4o, Anthropic Claude Sonnet, local Ollama
Recommended: Anthropic Claude Sonnet for email drafting — better instruction following and writing quality for business communication than GPT-4o in blind tests. Use streaming (SSE) to show draft appearing word-by-word.
Job queue
Inngest, BullMQ + Redis, Vercel Cron
Recommended: Inngest for snooze jobs, follow-up reminders, and background email sync. Serverless, works on Vercel without managing Redis.
Complexity estimate
Complexity 7/10 — the local-first architecture with SQLite caching and the Gmail/Graph API integration are the hardest parts. The AI features are straightforward. The native macOS and iOS apps add significant scope — plan for 4–6 months with a team of 3 for the web + desktop client.
Superhuman vs building your own
Open-source Superhuman alternatives
Existing projects you can self-host or use as a starting point. Each has trade-offs.
Mailspring
17.5KMailspring is an open-source email client (GPLv3) built with Electron, React, and TypeScript. Version 1.21.1 was released in May 2026. It supports Gmail, Outlook, iCloud, Yahoo, and generic IMAP — the provider breadth that Superhuman lacks. Mailspring includes read receipts, link tracking, email scheduling, and a plugin system. The architecture (Electron + React + local SQLite cache) closely mirrors what a Superhuman alternative would require.
Thunderbird (Android)
9KThunderbird for Android (MPL-2.0, formerly K-9 Mail) is Mozilla's open-source email client with IMAP, POP3, and Exchange support. The Android GitHub repository has 9K+ stars. The desktop Thunderbird project is separate. Thunderbird supports virtually any email provider, making it the broadest provider-support option.
Build vs buy: the real math
4–6 months
Custom build time
$120K–$280K
One-time investment
4–8 years for pure replacement; 12–18 months for vertical AI-email SaaS
Breakeven vs Superhuman
At 10 seats on Superhuman Starter, the annual cost is $3,000/yr. A custom build at $120K–$280K breaks even in 40–93 years on cost alone — Superhuman is not expensive enough to justify cloning as a cost reduction. The real case for building is verticalization: a vertical AI-email tool for a specific profession (legal, medical, sales, executive assistants) can charge $30–150/user/mo for domain-specific AI features that Superhuman doesn't build. An agency-focused email product with shared inbox, client labeling, and automatic CRM sync can charge $50–100/seat and break even at 30–50 seats in 12–24 months.
DIY roadmap: build it yourself
This roadmap covers building a vertical AI-powered email client using Mailspring's architecture as a reference. It assumes a team of 3 developers targeting a specific professional vertical (sales, legal, executive, or support teams). The focus is vertical AI features, not replicating Superhuman's brand.
Email integration and local cache
5–7 weeks- Implement Google OAuth 2.0 and Gmail API integration — read/write threads, labels, drafts
- Implement Microsoft OAuth and Graph API for Outlook/Exchange
- Build local SQLite cache with Dexie.js (web) or better-sqlite3 (desktop) for 5,000 recent emails
- Create background sync service: poll Gmail/Graph for new emails, update local cache
- Set up Next.js or React + Vite web client with keyboard shortcut system (hotkeys-js)
- Build inbox view rendering from local cache with real-time update overlay
Core email client features
4–6 weeks- Build compose window with rich text editor, keyboard shortcuts, and inline AI draft generation
- Implement split inbox using Gmail categories and sender-history classification
- Add snooze and follow-up reminder system using Inngest scheduled jobs
- Implement read receipt tracking via 1x1 pixel image embedding
- Build command palette (Cmd+K) with fuzzy search across all email actions
- Add email search against local cache with fallback to Gmail search API
Vertical AI features and billing
3–5 weeks- Build few-shot AI draft generation with user's sent history as style examples
- Implement vertical-specific AI features (e.g., legal contract clause extraction, sales objection detection, medical terminology recognition)
- Add CRM integration layer with HubSpot and Salesforce via OAuth
- Build shared inbox feature: multiple users managing a single email address
- Set up Stripe billing with per-seat workspace subscriptions
- Deploy Electron desktop wrapper for macOS using electron-builder
The 'feel of speed' that Superhuman is known for is hard to replicate — it requires the local SQLite cache architecture, careful keyboard shortcut implementation with no perceptible latency, and significant UX polish. Budget 20–30% of development time on performance tuning and UX refinement. The native desktop app adds 4–6 weeks if macOS-specific performance is a requirement.
Features you can't get from Superhuman
This is where a custom build pulls ahead — features impossible or impractical on a shared platform.
Shared inbox for support and customer success teams
Build a shared inbox where multiple team members see, claim, and respond to emails sent to a single address (support@company.com, success@company.com). Superhuman has zero shared inbox functionality. Implement as a virtual inbox layer that routes incoming emails to a shared queue in PostgreSQL, with assignment, collision detection, and internal notes separate from email replies.
Domain-specific AI with industry knowledge
Fine-tune or build RAG-powered AI drafting that understands your specific industry — legal contracts, medical correspondence, financial terms, or sales methodology. Superhuman's AI is generic. A legal email tool can recognize contract clause references, check tone against firm guidelines, and suggest appropriate legal language. Include a company knowledge base as RAG context for the AI model.
Automatic CRM logging and pipeline updates
Detect email conversations about deals, contacts, or support tickets and automatically log them to the CRM with correct associations. Superhuman's HubSpot/Salesforce integration requires manual syncing on Business tier. A custom build can use Claude to parse email content, identify deal stage implications, and push structured updates to Salesforce or HubSpot without any user action.
Email analytics dashboard for team managers
Track response times, email volume by client or deal, average response latency, and thread resolution rates across the team. Superhuman provides no team analytics visibility. A manager dashboard showing which team members are responding quickly and which have overdue threads creates accountability without requiring individual task management.
Universal IMAP support for non-Gmail users
Support Fastmail, iCloud, ProtonMail, and any IMAP-compliant email provider — the market that Superhuman explicitly ignores. Implement IMAP via the node-imap library or integrate Nylas's unified email API for broader provider coverage. This opens the product to privacy-conscious users, European teams (Proton, Tutanota), and organizations with custom mail servers.
Who should build a custom Superhuman
Sales teams needing CRM-integrated email intelligence
Sales teams pay $30/user/mo for Superhuman's speed while also paying $75–150/user/mo for Salesforce. A custom AI email tool with native bidirectional Salesforce sync, automatic email-to-deal logging, and AI drafts trained on successful sales emails delivers more sales-specific value at lower combined cost.
Privacy-conscious teams using non-Gmail providers
Organizations using Fastmail, Proton, Tutanota, or self-hosted email cannot use Superhuman at all. This is a sizable market — companies prioritizing email privacy or GDPR data residency. A custom AI email client with IMAP support serves this market that Superhuman has explicitly abandoned.
Support teams needing shared inbox with AI
Support teams using tools like Help Scout ($50/user/mo) or Intercom ($74+/user/mo) for shared inbox need AI-powered response drafting. A custom shared inbox email tool with Claude-powered draft generation covers both use cases — shared inbox and AI email — at a combined cost lower than separate Superhuman + Help Scout subscriptions.
Companies concerned about post-Grammarly roadmap
Grammarly's acquisition of Superhuman in July 2025 and the October 2025 parent rebrand signal that Superhuman's roadmap is now driven by Grammarly's writing AI strategy. Teams that adopted Superhuman for its email-specific features may find the product evolving toward a Grammarly writing assistant embedded in email — not what they paid for.
Skip the DIY — let RapidDev build it
Everything above is doable — but it takes months of full-time work. We build custom Superhuman alternatives using AI-accelerated development, delivering in weeks what used to take quarters.
Discovery call (free)
30 minWe map your exact requirements: which Superhuman features you need, what custom features to add, your users, integrations, and compliance needs. You get a detailed scope document and fixed-price quote within 48 hours.
AI-accelerated build
4–6 monthsOur engineers use Claude Code, Lovable, and custom AI tooling to build 3–5x faster than traditional development. You see progress in a staging environment every week — not a black box for months.
Launch + handoff
1 weekWe deploy to your infrastructure, transfer the GitHub repo, set up CI/CD, and walk your team through the codebase. You own 100% of the source code — no vendor lock-in, no recurring platform fees.
What you get
Timeline
4–6 months
Investment
$120K–$280K
vs Superhuman
ROI in 4–8 years for pure replacement; 12–18 months for vertical AI-email SaaS
30-min call. Fixed-price quote within 48 hours. No commitment.
Frequently asked questions
How much does it cost to build a Superhuman alternative?
Building a Superhuman alternative costs $120K–$280K for an MVP with Gmail/Outlook integration, local email caching, keyboard shortcuts, AI drafting, and snooze/follow-up features. Using Mailspring's open-source Electron + SQLite architecture as a foundation reduces scope by 4–6 weeks. The native macOS app adds $30–50K if required.
How long does it take to build a Superhuman clone?
4–6 months for a web + desktop client with Gmail/Outlook integration, AI drafting, and core productivity features with a team of 3. The Gmail API integration and local cache architecture take 5–7 weeks. The 'feel of speed' — sub-100ms keyboard interactions — requires an additional 2–3 weeks of performance tuning after the core features are built.
Are there open-source Superhuman alternatives?
Two notable options: Mailspring (~17.5K GitHub stars, GPLv3) is the strongest — an Electron + React email client with Gmail, Outlook, iCloud, and IMAP support, read receipts, and an active development community. Version 1.21.1 was released in May 2026. Thunderbird Android (~9K stars, MPL-2.0) offers the broadest provider support. Neither includes AI drafting out of the box.
Can RapidDev build a custom Superhuman alternative?
Yes — RapidDev has built 600+ applications including AI-powered productivity tools and email integration platforms. We recommend focusing on a specific vertical use case (sales, legal, support) rather than building a horizontal Superhuman clone. Visit rapidevelopers.com/contact for a free consultation.
Does Superhuman support IMAP or non-Gmail providers?
No — Superhuman officially supports only Gmail (personal and Google Workspace) and Microsoft Outlook/Exchange. There is no IMAP support, no Fastmail, no iCloud, no ProtonMail. This is a deliberate product decision that significantly limits Superhuman's addressable market and is cited in reviews as the primary reason teams consider alternatives.
What does Grammarly's acquisition mean for Superhuman users?
Grammarly acquired Superhuman on July 1, 2025 (TechCrunch). In October 2025, Grammarly renamed its parent company to 'Superhuman.' The strategic rationale is integrating Superhuman's AI email features with Grammarly's AI writing technology. For users, this means the product roadmap is now driven by Grammarly's broader AI writing strategy, not by email-specific power user needs. Features like shared inbox, IMAP support, and advanced keyboard customization may be deprioritized in favor of Grammarly AI integration features.
What is the hardest part of building a Superhuman alternative?
The local-first email caching architecture is the hardest engineering challenge — building a SQLite cache that synchronizes reliably with Gmail's API, handles incremental syncs, manages token refresh, and provides sub-100ms inbox loading requires careful architecture. The Gmail API's rate limits (250 quota units per second per user) and partial response patterns require intelligent caching to avoid hitting limits during full inbox sync.
Can I build a shared inbox on a Superhuman alternative?
Yes — and this is one of the strongest differentiators since Superhuman offers no shared inbox on any plan. Implement shared inbox as a virtual mailbox layer: incoming emails to the shared address are stored in a shared queue in PostgreSQL, with assignment rules, collision detection (prevent two agents from replying simultaneously), internal notes, and SLA tracking. This covers the use case that drives teams to Help Scout or Intercom.
We'll build your Superhuman
- Delivered in 4–6 months
- You own 100% of the code
- No per-seat fees, ever
30-min call. No commitment.