Chatbot Open-Source: What Support Teams Should Know
A chatbot open-source setup is software whose original code is available for people to inspect and modify under its licence. For customer support, the code is only one part of the decision: your model, knowledge, interface, hosting and human handoff all affect what customers experience.
The chatbot’s code is available to inspect and modify under its licence. Other parts of the system may still be proprietary.
Choosing the right setup
Chatbot open-source software makes its original code available for inspection and modification under the terms of a licence. That can give a team room to adapt how a bot works, but “open source” does not automatically mean free to operate, easy to maintain, or entirely open. A chatbot may combine an open framework with a proprietary model, hosted interface, paid service, or separate connector. Treat these as separate choices, not one label.
What must be open, and what does it cost to run?
A chatbot open-source project may publish its framework while relying on separate models, connectors or hosted services. Check each component’s licence and terms before treating the whole setup as open source.
Before adopting a tool, map the layers: the model generates language; orchestration routes the conversation; retrieval finds useful passages in your help content; the interface lets customers ask questions; connectors link other systems; and deployment determines where the software runs. Each layer can have its own licence and operating requirements. A public codebase alone does not tell you whether the model, interface, connectors, or hosting are open or included.
A self-hosted setup can put more deployment decisions in your team’s hands, but your budget should include more than the software licence. If you are comparing a self-hosted help desk, account for the people and systems needed to operate it as well. Allow for infrastructure, any model or API usage, engineering and integration work, updates, monitoring, and keeping policies current. The workload depends on the system and how you use it; do not assume that an open-source download has no continuing cost. Likewise, a hosted plan fee and a self-hosted operating budget are different kinds of costs.
For a support team, test the whole service rather than a polished demo. Teams comparing a live chat software open source setup should also check how it handles questions that need a human. Can the bot find the current returns policy? Does the customer see which content informed the answer? What happens when the policy does not cover an exception? A useful system needs a clear route to a person as well as a way to answer routine questions. If self-hosting is part of your decision, use this guide to self-hosted AI chatbots to think through setup and operating trade-offs.
When source-available support software is a fit
Open source, source-available and self-hosted describe different things. Check whether a licence permits your intended commercial use, modifications, hosting for others and redistribution; then decide whether you want to operate the software yourself or use a hosted support workflow.
Know what the licence permits
momo is source available under the PolyForm Internal Use License 1.0.0: a company can run and modify it for its own use, but cannot resell it, host it for others or redistribute it.
The licence gives a clear boundary between internal use and providing the software to others.
Ground answers in your content
It can use a website crawl, PDFs, Word files, plain text and Q&A pairs as knowledge, then send a confident answer with citations when its source check supports the reply.
Support replies can draw on the policies and product information your team provides.
Keep a person in the conversation
When the AI is not confident, it tells the visitor it is not sure and opens a ticket in the shared helpdesk inbox, where a teammate can take over.
Questions that need judgement reach the team rather than being treated as routine answers.
Compare operating costs with hosted plan fees
A self-hosted chatbot budget can include infrastructure, model or API usage, engineering, integration, maintenance and monitoring; hosted plans instead have a plan fee and included AI conversations.
Included: Flat plan fee with an included number of AI conversations; no per-resolution fee or credit packs. · The helpdesk inbox is included, and teammates can take over conversations live. · An AI conversation can contain up to 25 AI replies on paid plans and up to 10 on Free. · Plans are sold in euros; yearly billing is 11 times the monthly price.
Free
Try momo with real visitors.
€0€0 forever
- 30 AI conversations, one-time
- 1 agent, 1 team member
- Website widget
- Helpdesk inbox
- Basic analytics
- 2 MB of knowledge, 10 sources
Starter
One agent, steady traffic.
€29€27per month
- 500 AI conversations / month
- 1 agent, 2 team members
- Website widget and email
- Helpdesk inbox and Help Center
- Basic analytics
- No momo badge
- 20 MB of knowledge, 50 sources
Growth
More agents, more channels.
€129€118per month
- 2,000 AI conversations / month
- 3 agents, 5 team members
- Slack, Shopify, WhatsApp, Instagram, Messenger
- API, Zapier and webhooks
- Advanced analytics
- 100 MB of knowledge, 200 sources
Scale
High volume, every channel.
€399€366per month
- 6,000 AI conversations / month
- 10 agents, 10 team members
- Every channel
- API, Zapier and webhooks
- Advanced analytics
- 500 MB of knowledge, 1,000 sources
Need more resources or custom solutions? Contact us for Enterprise plans
Chatbot open-source questions
Use these checks to separate licence terms from the other decisions involved in customer support.
Is a chatbot open-source if only its framework code is public?
Not necessarily. A chatbot can combine a public framework with a proprietary model, hosted interface or connectors, so check the licence and documentation for each layer. Publicly available code does not by itself make the complete service open source.
Can an open-source chatbot use a proprietary AI model or API?
Yes, a chatbot can combine open components with a proprietary model or API. The model, orchestration, retrieval, interface and deployment are separate parts of the system, and each may have different terms. Check what data is sent to any external service before using it for customer conversations.
Can you self-host a chatbot, and how much engineering does it need?
Hosting options and licence terms can change, so verify current official documentation before choosing a tool. Requirements also depend on the deployment, integrations and support workflow you intend to build. Do not use an old self-hosting description as proof of current availability.
What does it cost to run a self-hosted chatbot beyond the software licence?
Plan for infrastructure, model or API usage, engineering, integrations, updates, monitoring and knowledge upkeep. The actual budget depends on your deployment and workload, so a licence price alone is not a total cost estimate. Compare those ongoing responsibilities with a hosted plan’s fee and included usage.
How can a chatbot answer from a returns policy and hand uncertain cases to a human?
Make the current policy available to the chatbot, ask a specific return question, and check the reply against the retrieved passage. Then ask about an exception the policy does not settle and inspect whether the system routes it to a person instead of guessing. For momo, a reply that does not meet its confidence line is treated as uncertain and opens a ticket for the team.
What should I check in a chatbot licence before using it for commercial support?
Read the terms for commercial use, modification, hosting for other organisations and redistribution. Also check whether models, connectors and hosted services have separate terms. momo is source available under the PolyForm Internal Use License 1.0.0; it can be run and modified for a company’s own use, but not resold, hosted for others or redistributed.
Start with one policy question
Give your team a clearer way to handle policy questions, with cited answers and a route to human support when a policy leaves an issue unresolved.