Intercom invented modern in-app messaging and remains the most complete customer communication platform in the business: live chat, help desk, product tours, outbound campaigns, and now Fin — an AI agent that genuinely resolves a meaningful share of support tickets on its own.
It is also famously, structurally expensive: $39+ per seat per month on the plans that matter, plus usage-based pricing for Fin resolutions, plus add-ons — small teams routinely report bills of several hundred dollars a month. If what your business actually needs is "a human answers chat on our website", you are buying a platform to use a bubble. Chatlet is the bubble, done properly, for $49 once. Here is the honest comparison.
What Intercom does well
Intercom is the deepest product in the category:
- Fin AI agent — genuinely state-of-the-art automated resolution, not a gimmick.
- The full platform: help desk, knowledge base, product tours, outbound messages and campaign targeting.
- Rich customer profiles with event tracking, segmentation and workflow automation.
- Enterprise-grade integrations, reporting and mobile agent apps.
If support is a funded department and AI deflection at scale moves your unit economics, Intercom is priced against real savings.
Where the subscription model hurts
The pricing model compounds three ways: per seat (every agent added is $39–99/month), per usage (Fin bills ~$0.99 per resolution), and per module (tours, campaigns and surveys are add-ons). Teams describe the same arc — start small, watch the bill triple as seats and usage grow, then discover downgrading means losing workflows you built around. Three years for even a two-seat setup lands in the thousands.
For a small SaaS, an agency, or an e-commerce shop, the honest question is: which of those platform features are you using? If the answer is "we answer chats when customers ask questions", then per-seat AI-platform pricing is the wrong shape entirely — you are renting a contact center to run a doorbell.
Chatlet: the pay-once alternative
Chatlet is a $49, one-time purchase. The live-chat bubble you own — unlimited agents, unlimited sites, no per-seat pricing. Chatlet is deliberately the anti-platform: one small Node server, one SQLite file, a shadow-DOM widget with WebSocket delivery and polling fallback, an agent dashboard with typing indicators, visitor context and canned /shortcuts, offline messages with email notification via your own SMTP, unlimited agents, unlimited sites, transcripts on demand — and a price that never recurs.
The source code is MIT-licensed at github.com/bensblueprints/live-chat — free to build and run yourself, forever. Buying the packaged version on Whop gets you the signed installer, 1-click setup and updates. Either way, there is no account, no telemetry and no renewal date.
Head to head
| Chatlet | Intercom | |
|---|---|---|
| Price | $49 once | $39+/seat/mo + usage-based AI fees |
| Cost over 3 years | $49 | ~$1,404+ per seat |
| Where your data lives | Your machine / your server | Their cloud |
| Usage limits | None — unlimited use | Per-seat + per-resolution + add-on modules |
| Works offline | Yes | No |
| Source code | MIT, on GitHub | Closed |
Who should stay with Intercom
Stay with Intercom if Fin's automated resolutions demonstrably pay for themselves, if you run genuine outbound campaigns and product tours, or if your support org needs enterprise workflows, SLAs and reporting. Chatlet competes with none of that.
Switch if your support volume is human-scale and your need is human chat: Chatlet gives every agent a seat for free (there are no seats), every site a widget for free (there are no per-site fees), and your customer conversations a home on your own server.
Making the switch
Step 1 — Buy once on Whop. One-time $49 for the packaged version — Windows desktop app plus one-command VPS deploy.
Step 2 — Deploy where visitors can reach it. docker compose up -d on a $5 VPS behind your reverse proxy (WebSocket upgrade passed through). Desktop mode is perfect as the always-on agent console.
Step 3 — Paste the snippet, start chatting. Copy the embed from Sites, paste it before