Conversational Automation · SaaS 2026 In development

Kineto Visual Bot Builder

A visual, multi-tenant platform for building and publishing conversational workflows across WhatsApp and web, with AI and business-tool integrations built in.

2
Chat channels
Visual
Flow builder
3
Integration surfaces
Stack Next.js 16TypeScriptSupabaseReact FlowOpenAIAnthropicZustandReact Query

The Problem

Building a production conversational bot usually means stitching together a flow editor, channel-specific message handling, session storage, AI providers, business integrations, credential security, and deployment tooling. Most solutions either lock teams into a rigid template or scatter that logic across separate services that become difficult to test, version, and maintain.

Kineto needed to give non-technical teams a visual building experience without sacrificing the execution controls, security boundaries, and extensibility required by engineering teams. The same workflow also needed to behave consistently across WhatsApp and web while supporting external tools and long-running conversations.

The Product

Kineto is a visual bot-building platform for designing conversational automations without hard-coding every interaction. Teams compose flows on a node-based canvas, test them in a web preview, and publish versioned bots to WhatsApp and web.

The platform goes beyond a simple chatbot editor. A centralized execution engine handles branching logic, variables, validation, integrations, session state, and human-facing messages as one portable workflow.

What I Built

Visual Workflow Editor

The editor uses React Flow to turn messages, inputs, conditions, scripts, and integrations into connected blocks. Zustand manages responsive canvas state, while React Query owns server-backed workspace and bot data.

Multi-tenant SaaS Foundation

Kineto supports workspaces, role-based memberships, folders, bot settings, working drafts, and immutable published versions. Supabase provides Postgres, authentication, storage, and row-level security.

Channel-independent Execution

The chat engine normalizes web and WhatsApp messages into the same internal format. The flow runner can pause for external webhooks, resume a session, interpolate variables, validate input, and deliver the next set of outputs without coupling business logic to a single channel.

Production Integrations

The builder includes OpenAI, Anthropic, Google Sheets, HTTP, email, Cal.com, and WhatsApp building blocks. Workspace credentials are encrypted, and companion apps expose Kineto workflows through Make.com and Zapier.

Architecture

  • Next.js 16 App Router for the application and API boundary
  • TypeScript contracts shared across the editor and execution engine
  • Supabase Postgres, Auth, Storage, and row-level security
  • React Flow for the node canvas
  • Zustand and React Query for client and async state
  • Isolated browser and server runtimes for user-authored script blocks
  • Nx and pnpm for the web, sandbox, Make.com, and Zapier monorepo

The Result

Kineto brings workflow design, live preview, versioned publishing, session persistence, and third-party automation into one cohesive product. The architecture keeps the core flow portable so new channels and integration blocks can be added without rebuilding the execution model.

← Previous Multi-Agent Patient Intake Workflow Next → AI Wireframe Generator

Want something
like this?

30 minutes, free, no deck. We'll figure out if I'm the right fit for your project.