Intercom charges $74 per seat per month on the Essentials plan. Zendesk Suite starts at $55 per agent per month. For a five-person support team, that is $3,300 to $4,440 per year before any add-ons. Chatwoot does the same job for free if you run it yourself, or $19 per agent per month on the cloud plan. For a support ops manager who has started asking why a software-as-a-service vendor controls their customer conversation history, or a founder watching per-seat costs climb as the team grows, Chatwoot is a direct comparison.
| Fact | Value |
|---|---|
| What it is | Open-source customer support platform with omnichannel inbox, live chat, email, WhatsApp, and AI assistant |
| Built by | Chatwoot, Inc. with open-source contributors |
| License | MIT license (a permissive open-source license that lets you run, modify, and redistribute the software without restriction) |
| Platforms | Web app (agents), embeddable chat widget (customers) |
| Requires | Ubuntu 20.04 or later, PostgreSQL 12+, Redis 6+, 2 CPU cores, 4 GB RAM |
| Install method | Linux script, Docker Compose, Kubernetes, or one-click deploy on DigitalOcean and AWS |
| Verdict | Install if you are paying Intercom or Zendesk per-seat fees, or if your team needs to keep support conversation data on your own infrastructure |
What Chatwoot actually is
Chatwoot is an open-source customer support platform that consolidates customer conversations from multiple channels into a single inbox. The idea is simple: instead of having your team juggle a separate tab for live chat, another for email support, and a third for WhatsApp messages, everything lands in one place for any agent to pick up and reply to.
This concept is called omnichannel support, meaning a single inbox that collects messages from multiple channels including live chat, email, WhatsApp, Facebook, Instagram, Telegram, and others, all in one place. Chatwoot is not the only tool that does this, but it is the most established open-source project taking on Intercom (intercom.com) and Zendesk directly.
The codebase lives at github.com/chatwoot/chatwoot under an MIT license. The MIT license is a permissive open-source license: you can run it, modify it, embed it in a commercial product, and distribute it without any obligation to publish your changes. That is a meaningfully different licensing posture from tools like Cal.com, which use AGPL (a copyleft license requiring you to share modifications), and it matters if you are building a product on top of the support platform.
The project has 34,900 GitHub stars, 8,400 forks, and over 6,500 commits. It has been actively developed since 2019.
What you get after installing it
Inbox management and channels
- Live chat widget that embeds on your website with a few lines of JavaScript
- Email support inbox that routes incoming support emails directly into Chatwoot
- WhatsApp Business API and Twilio-powered SMS for messaging channels
- Facebook Messenger, Instagram, and Telegram integrations
- A single unified queue where all of these land, assigned to agents or teams based on rules you configure
Agent and team tools
- Canned responses: pre-written replies that agents can insert with a shortcut, cutting handle time on common questions
- Private notes on conversations, visible only to your team, useful for internal context without the customer seeing it
- @mentions to loop in a colleague on a conversation thread
- Labels and custom attributes for tagging and filtering conversations by product, issue type, priority, or anything else
- Team-based conversation routing so an e-commerce team and a technical team can each see only their relevant queue
Customer context
- Contact profiles that aggregate a customer's conversation history, custom attributes, and any data passed in via the API (a connection interface that lets other software plug into Chatwoot)
- Contact segmentation so you can filter contacts by attributes, useful for seeing all users on a specific plan or all customers in a specific region
Automation and AI
- Workflow automation that triggers actions (assign conversation, send message, add label, set status) when conditions are met. This is built without code through a visual rule builder.
- Webhooks (automatic notifications sent to another tool when something happens in Chatwoot) so you can push conversation events to your CRM, data warehouse, or any other system
- Captain, Chatwoot's AI assistant, which can read your help center articles and past conversations to suggest replies and deflect common questions before they reach an agent. Captain is included on paid plans and available on self-hosted.
Reporting
- CSAT (customer satisfaction score, a rating customers give after a support conversation) reports by agent, team, or label
- Conversation volume reports by channel, date range, and assignment
- Agent response time and resolution metrics
Help Center
- A built-in knowledge base where you can publish articles and FAQs. Customers can search it from the live chat widget before opening a conversation, which reduces inbound volume for questions you have already answered.
The self-hosted version gives you all of these features. Captain AI credits for the AI features are usage-based; the cloud plans bundle a monthly credit allocation starting at 300 credits per agent on the Startups tier.
The install experience
The two practical paths are the automated Linux script and Docker Compose.
The Linux script is the fastest route if you are deploying to a fresh Ubuntu server. You run a single command from the terminal, follow a few prompts for your domain name and SMTP (the email delivery protocol used to send notification messages) settings, and the script handles the rest: installs dependencies, configures PostgreSQL and Redis, sets up Nginx (a web server that sits in front of Chatwoot and handles incoming traffic) as a reverse proxy (a layer that forwards requests from the internet to the right internal service), and issues an SSL certificate via Let's Encrypt (a free service that encrypts traffic between your server and your users' browsers). Most people complete this in under 30 minutes.
Docker Compose is the right choice if you want more control over the deployment. Docker is software that packages an application and all its dependencies into a container, so it runs identically on any server. Docker Compose is the tool that coordinates multiple containers together (Chatwoot's app, its database, and its cache all run as separate containers). This approach suits teams that plan to run Chatwoot alongside other containerized services, or are deploying to a platform like DigitalOcean's App Platform. The steps are: download the environment file and Docker Compose configuration from the repository, set your database and Redis passwords in the environment file, initialize the database schema, and start the containers with docker compose up -d. After the containers are running, you set up Nginx and SSL. The process takes about the same amount of time as the Linux script but requires more manual configuration decisions.
The honest difficulty assessment: if you have deployed a web application before, either path is routine. If you have not, the Linux script is workable with the documentation, but expect to spend time reading and troubleshooting, not just clicking. The minimum server specification is 2 CPU cores and 4 GB RAM, which maps to roughly a $12-20 per month VPS (a rented slice of a remote machine) from DigitalOcean, Hetzner, or Vultr.
One-click deploy options exist for DigitalOcean Marketplace, AWS, and Google Cloud. These handle the infrastructure provisioning and reduce setup to filling in configuration values, with the trade-off that the underlying infrastructure cost is determined by the cloud provider rather than by you.
If you want to avoid infrastructure work entirely, the cloud plan starts at $19 per agent per month on the Startups tier with unlimited conversations and all channels.
Where it fits and what to compare it to
Intercom (intercom.com) is the default comparison. The product quality is high, the AI features are mature, and the integrations are broad. The gaps: pricing starts at $74 per seat per month on Essentials and escalates from there, there is no self-hosting option, and your customer conversation data lives on Intercom's infrastructure. If your team has 10 agents, that is $8,880 per year before add-ons.
Zendesk is the other natural comparison. It covers the same channel breadth as Chatwoot, has a more mature ticketing system for teams managing high-volume support, and is priced starting at $55 per agent per month on Suite Team. Like Intercom, no self-hosting. For teams that need enterprise-grade SLAs (service-level agreements, the contractual commitments a vendor makes about uptime and response time) and compliance built in before they configure anything, Zendesk is a reasonable choice. For teams where the per-seat cost is the main constraint, Chatwoot is a direct substitute.
Zammad is another open-source support platform worth knowing. It is Ruby-based like Chatwoot, has a comparable channel set, and uses an AGPL license (a copyleft open-source license that requires you to publish any changes you make to the code) rather than the more permissive MIT. The interface is more structured around ticketing workflows than live chat. Zammad is worth evaluating if your team's primary channel is email tickets rather than live chat.
Crisp (crisp.chat) is a closed-source live chat and support platform with a free tier for two agents. It covers live chat, email, and some messaging channels, and is easier to get started with than self-hosted Chatwoot. The trade-off is no self-hosting and data residency on Crisp's servers.
Chatwoot's specific case: you are running a support team of three or more agents, you are paying Intercom or Zendesk per-seat fees that compound as you grow, you need your customer conversation history in a database your organization controls, or you are building a product where support is embedded and you want the flexibility of an MIT-licensed codebase. On those scenarios, Chatwoot earns a direct look. For a solo operator or a two-person team with no infrastructure capacity, the cloud free tier (up to two agents) or Crisp's free plan is the lower-friction starting point.
Verdict
Install, with the usual self-hosting caveats.
The cost argument is blunt. If you have four agents on Intercom Essentials, you are paying around $3,552 per year. The same team on Chatwoot's cloud Startups plan costs $912 per year. Self-hosted on a $20 per month VPS brings it to $240 per year. The feature surface is comparable on everything except Intercom's highest-tier AI capabilities, which require separate credits on both platforms anyway.
The MIT license is worth naming as a practical advantage. You can modify the codebase, embed it in a product you sell, and run it indefinitely without owing anything to anyone. That is a meaningful difference from AGPL-licensed tools and from proprietary alternatives where your access depends on a vendor relationship.
The caveats: Captain AI features require cloud credits and are not unlimited on self-hosted deployments without configuration. The install path requires technical comfort with Linux servers or Docker. There is no managed update path built into the self-hosted version; you are responsible for applying version updates yourself, which is standard for self-hosted open-source software.
For a support ops manager, customer success lead, or founder who is currently writing a check to Intercom or Zendesk every month, Chatwoot is the clearest self-hosted alternative on the market. The data stays in your database, the per-seat cost is a fraction of the incumbents, and the project has the kind of GitHub activity (34,900 stars, active commits in 2026) that signals real-world adoption and continued maintenance.