Suppa

The editor is a node canvas, and the node set is practical rather than decorative. A start point triggers on user consent, a branch node routes on conditions, and a user input node stores replies in a variable you can reuse later. You can wire a send message node for fixed copy, a knowledgebase node for retrieved answers, or an AI response node when you want the model to write the reply itself. Flows stay readable because each node does one job.

Suppa takes content from multiple sources and uses it as the grounding layer for AI answers. Each plan includes a fixed number of knowledgebase items, from five on Starter to fifty on Business, so the practical question is how tightly you can scope your content. A tight scope is a feature, not a limitation: a bot with twenty clean pages about your product answers better than one stuffed with the whole intranet. The changelog also notes sub-page crawling for URLs, which matters when your help center is spread across nested pages.

An API call node hits your own endpoint mid-conversation. That's what makes the bot useful beyond FAQ duty. A customer can ask where an order is, the flow calls your order service, and the answer comes back with a real status and date. Each plan allows a set number of API actions, five on Starter and one hundred on Business, so heavy backend use pushes you up a plan. Third-party services work the same way if they expose an endpoint or a Zapier hook.

You pick the language model behind the assistant. The current lineup covers GPT-4o Mini, GPT-4o, Claude 3 Haiku, and Claude 3.5 Sonnet. Mixing a fast cheap model for routing with a stronger one for written answers is the obvious play, and Suppa leaves that decision to you instead of locking one provider. Model quality matters most on the free-form paths, so put your best model where the bot writes, not where it sorts.

The channel list is wider than most tools at this price. You can embed a widget on a website, install the Android or iOS SDK, or connect WhatsApp, Telegram, and Instagram. A WhatsApp chatbot reaches customers where they already message businesses, which is often the difference between a bot people use and a bot people ignore. The whitelabel API covers setups where the assistant has to appear as part of your own product.

The dashboard reports on conversations, and higher plans add advanced analytics, custom reporting, and weekly autotraining. Autotraining is the one to watch: it means the knowledge layer gets refreshed on a schedule instead of drifting out of date. Automatic lead collection runs alongside the chat, so a support conversation can also produce a sales lead. Basic analytics and email support come with every plan, including Starter.