Let me start with the uncomfortable part: Intercom is good software. The messenger is polished, Fin is a genuinely capable AI agent, and the whole thing feels like one product rather than a bundle of acquisitions. If your support volume is high and your budget is elastic, you can stop reading here.
People do not search for "open source Intercom alternative" because Intercom is bad. They search because the bill grows in three directions at once (seats, resolutions, and usage), or because their customer conversations live in someone else's database. Sometimes the trigger is simpler: procurement asked what this costs next year, and nobody could answer. We wrote up the full billing mechanics in our Intercom pricing teardown; this post is about what replacing it looks like in practice.
What "replacing Intercom" actually means
Most teams use a fraction of Intercom. In practice the workload breaks down into four jobs:
- A messenger on your site or app where visitors and customers start conversations.
- A shared inbox where your team answers them, with assignment, tags, snoozing and canned responses.
- A help center that deflects the questions you have already answered ten times.
- Outbound messaging: onboarding emails, announcements, and the transactional mail your product sends anyway.
Any alternative has to cover those four. Messy does, and adds the channels Intercom treats as paid add-ons (SMS, WhatsApp, mobile and web push) as first-class citizens behind one API.
Feature by feature
| Job | Intercom | Messy |
|---|---|---|
| Chat widget / messenger | ||
| Shared inbox with assignment, tags, snooze | ||
| Help center | ||
| Email campaigns and drip sequences | ||
| Transactional email, SMS, WhatsApp, push via API | Add-ons and usage fees | Included, bring your own provider keys |
| AI agent answering tickets autonomously | Fin, $0.99 per resolution | |
| Product tours and in-app surveys | ||
| Self-hosting, your database | Free, Elastic License 2.0 | |
| Pricing model | Per seat + per resolution + usage | Free self-hosted, or flat $20/mo cloud |
The two honest gaps are Fin and product tours. If an AI agent currently resolves half your ticket volume, that is real value and you should price it in before switching. If you mostly use Intercom as a shared inbox with a widget in front of it, you are paying enterprise prices for a queue.
The math at 5 and 20 seats
Listed prices as of this writing: Intercom's mid tier runs $85 per seat per month billed annually, and Fin bills $0.99 for every conversation it resolves. Take a 5-person support team handling a modest 400 Fin resolutions a month:
- Intercom: 5 × $85 + 400 × $0.99 ≈ $821/month, before email volume and channel usage.
- Messy cloud: $20/month flat, unlimited seats, plus what you pay Amazon SES or Twilio directly (SES is roughly $0.10 per thousand emails).
- Messy self-hosted: the cost of a small VM.
At 20 seats the Intercom line crosses $2,000 a month while the Messy line does not move. Unlimited seats is not a promotional bullet point; it falls out of the model. When you self-host, there is nobody to meter your logins. The cloud plans keep the same principle because charging per seat punishes you for involving your whole team in customer conversations, which is exactly what you want to encourage.
What switching looks like
A migration is smaller than it feels. The messenger is a script tag swap. The inbox is where your team lives, so give it a week of parallel running: point a test page at the Messy widget, let new conversations land in the shared inbox, and keep Intercom open for the old ones until they drain.
- Contacts: export from Intercom as CSV, import into Contacts. Custom attributes come along and power segments on the other side.
- Saved replies: recreate them as canned responses. This is an afternoon of copy-paste, or one prompt if you use the playbooks approach with a coding agent.
- Help center articles: port the top twenty by traffic first. The long tail can wait.
- Outbound: rebuild active sequences as drips. Most teams discover half their Intercom series were switched off anyway.
Who should not switch
Stay on Intercom if Fin resolves a large share of your volume profitably, if you lean on product tours and surveys as a core growth motion, or if you have nobody comfortable running software (in that case at least use a hosted plan rather than self-hosting). Switch if your Intercom invoice has become a quarterly negotiation, if data residency matters to you, or if you want your messaging stack to be something you own rather than rent.
The source is on GitHub under the Elastic License 2.0. The fastest way to evaluate it is to run it: our self-hosting guide covers what a production deployment involves.
