Your no-code chatbot builder.

A no-code chatbot builder lets you connect messages, data collection and routing without writing the conversation logic in code. Gleap adds Kai Support for answers from your product knowledge, and routed tickets reach your team in the shared inbox.

Yes. Start a workflow on new conversations, add Buttons for your most common questions and a Let Kai answer step, and Kai Support answers frequently asked questions from your help center, snippets and files, with the sources listed. Questions it cannot answer go to your team through an Assign ticket step. Unlike a scripted FAQ bot, Kai Support writes each answer from the matching content, so newly published help articles are used without editing the flow. See Kai Support, the AI support agent. Shops use the same setup as an AI chatbot for ecommerce, with order lookups from Shopify or WooCommerce.

No. You build conversation workflows in the visual editor. Custom API actions and starting a workflow from your own application may need developer setup. The available action types are listed in the workflow action guide.

Supported triggers include conversation start, ticket creation, reopening or closure, status changes, customer or agent comments, and no-response conditions. You can also start a workflow manually through the SDK. Set the channel, audience, and priority for the intended use case.

Yes, but overlapping published workflows can compete. For triggers that select one workflow, the highest priority wins; equal priorities can produce an unpredictable choice. Review the editor’s conflict warnings and test the intended path.

A workflow follows the actions and conditions you define. A custom agent, on Team and up, uses instructions and configured tools for a task, and can also use an existing workflow as a tool. From Team, Kai Support itself can use tools your team adds, such as API calls and MCP servers. Choose based on the task, and keep access and consequential actions within your team’s intended scope.

Yes. The editor includes custom API and integration actions. Configure the connection and inspect the result for your specific system. An available integration does not mean every action in that system is automatically enabled. See the integration directory.

No. A workflow remains a draft until it is published. Run it in the messenger and inspect its execution before rollout.

Yes. Save and run the draft as a guest or selected contact, then inspect the execution timeline. The test creates a real conversation in your messenger, so use test data and check each branch.

Your team configures the endpoint, authentication and data used by a custom API action. Enable only the access the workflow needs and test the response before publishing.

Workflows are available on every Gleap plan. A Let Kai answer step bills like Kai Support, by the AI it uses. Escalating a conversation for Kai Support to investigate, or passing it to a custom agent, needs Team or higher. See Gleap pricing.

A support workflow can collect the context that starts a broader product improvement. Self-driving development connects customer signals to investigation, reviewable changes and follow-up.

No. A chatbot workflow defines conversation actions. Kai Code handles coding sessions and reviewable pull requests as a separate part of the customer feedback loop.