What Chatbase does well
Chatbase built its reputation on one thing: getting you from sign-up to a working chatbot faster than any competitor. Their onboarding wizard is genuinely polished — you enter a URL, watch the crawl progress bar, and within five minutes you're having a real conversation with a bot trained on your site. For someone who wants to evaluate AI chatbots quickly, that experience is hard to beat.
Beyond onboarding, Chatbase has some genuine strengths worth acknowledging:
- Channel integrations: Chatbase offers native integrations with Slack, WhatsApp, and Messenger, which matters if your customers reach you on those platforms rather than your website.
- UI themes: The default chat widget designs are clean and professional. Multiple pre-built themes mean you can get a good-looking widget without custom CSS.
- API access: On higher-tier plans, Chatbase exposes an API that lets you query your chatbot programmatically — useful if you want to embed answers inside an existing product UI rather than a floating widget.
- Broad adoption: Because Chatbase is one of the most widely used tools in this space, there's a large community of users, tutorials, and third-party guides. If you get stuck, someone has probably solved your problem already.
Where Chatbase falls short
Despite its strengths, Chatbase has a set of structural limitations that become significant at scale or when brand consistency matters:
Per-message pricing that compounds quickly. Chatbase's lower-tier plans charge per message or per conversation, which sounds cheap in the abstract but becomes painful at volume. A site with 2,000 monthly visitors where 10% engage with the chatbot at an average of 8 messages per conversation is 1,600 messages per month — already stressing the free tier and requiring a paid plan. As traffic grows, the bill grows unpredictably.
Limited white-labelling on entry-level plans. The "Powered by Chatbase" attribution in the widget is prominent on lower tiers. For a bootstrapped startup, this might be acceptable. For an established brand that takes presentation seriously, having a vendor's logo embedded in your customer experience feels unprofessional — and paying to remove it means upgrading to a significantly more expensive plan.
Data portability concerns. Your conversation history — which question was asked, how the bot responded, whether the user was satisfied — lives on Chatbase's infrastructure. Exporting that data requires specific plan tiers and uses their proprietary format. If you ever need to migrate away from Chatbase, reconstructing your conversation history elsewhere is non-trivial.
No built-in lead capture. When the bot can't answer a question, there's no native mechanism to capture the user's email for a human follow-up. This is a significant missed opportunity — unanswered bot questions are warm leads — and you'd need a custom integration to replicate it.
When to go fully custom
Building a chatbot entirely from scratch — your own vector database, your own embedding pipeline, your own UI, your own LLM API integration — is the right choice in a narrow set of circumstances:
- You have compliance requirements (HIPAA, FedRAMP, SOC 2 Type II) that prohibit using third-party managed services for any part of the pipeline.
- You need deep, real-time integration with your own product's database — the bot needs to query live account data, look up order status, or check inventory in real time, not just retrieve from static content.
- You have an in-house ML team that can maintain the infrastructure, and a product roadmap that justifies the investment.
- The chatbot is the product — you're building an AI assistant that will be a core differentiator in your product, not a support widget on the side.
Outside those scenarios, fully custom is almost always over-engineering. A competent team will spend 3–6 months building something that a managed tool provides out of the box today. The opportunity cost of that engineering time — features you didn't ship, customer problems you didn't solve — typically exceeds the marginal benefit of full control.
The middle ground: managed custom solutions
Between "use Chatbase as-is" and "build everything from scratch" is the category Sitepilot occupies: managed tools that provide the speed and reliability of a SaaS product with the control and flexibility that Chatbase lacks.
The defining characteristics of this category are flat monthly pricing (so costs don't scale with conversation volume), full white-labelling with your own branding, data portability (you own your conversation logs), and features like lead capture that treat the chatbot as part of your business workflow rather than an isolated support widget.
For most businesses — SaaS companies, e-commerce brands, service businesses, agencies — this middle ground is the right answer. You get to deploy in days rather than months, avoid the engineering overhead of custom infrastructure, and still maintain control over your brand and your data.
Decision matrix
| Your situation | Recommended approach |
|---|---|
| Low traffic (<500 conversations/month), just evaluating, no brand requirements | Chatbase free or starter tier — fast to try, low risk |
| Growing traffic (500–5,000 conversations/month), brand matters, want flat pricing | Managed custom solution (Sitepilot) — predictable cost, white-label, lead capture |
| High traffic (>5,000 conversations/month), need channel integrations (Slack, WhatsApp) | Chatbase Pro or Enterprise tier — their channel integrations are ahead of most competitors |
| Real-time product data queries, deep CRM integration on day 1, compliance requirements | Fully custom build — the complexity is justified by the requirements |
| Agency building for clients, need to white-label for multiple brands | Managed custom solution with white-label and multi-bot support |
Migration from Chatbase to Sitepilot
If you're currently on Chatbase and want to switch, the process is straightforward and typically takes under an hour:
- Export your conversation logs from Chatbase before cancelling. Download them in CSV format from your analytics dashboard. Even if you can't import them into Sitepilot directly, they're valuable for your records and for understanding your customers' most common questions.
- Create a new bot in Sitepilot and paste your website URL. The crawl typically completes in 2–5 minutes for sites under 50 pages.
- Recreate any manual Q&A pairs you had in Chatbase. If you added specific hardcoded answers to your Chatbase knowledge base, you'll need to re-add those in Sitepilot's knowledge base editor.
- Update your embed script. Remove the Chatbase script tag from your site and replace it with the Sitepilot script tag. This is a one-line change in your site's HTML or CMS.
- Test in incognito mode before going live. Ask your 10 most common questions and confirm the bot handles them correctly with the new setup.
The migration is non-destructive — you can run both embed scripts simultaneously for a day or two while you verify the new bot performs correctly, then remove the Chatbase script. There's no downtime and no risk of leaving your visitors without a chatbot during the transition.