AI-native SaaS, not AI-flavored boilerplate
A real product surface before you write your first feature
CentraKit starts where most kits stop: tenant workflows, AI-ready product surfaces, API and MCP access, and boundaries that help agents and engineers ship the next feature safely.
centrakit.app
Tenant workflow preview
Animated cards show how customer work, AI actions, and product state stay connected.
Live system
Intake
3Customer matched
Asset tagged
Quote drafted
In progress
3AI summary
Task assigned
SLA checked
Ready
3Approval sent
API synced
Docs updated
Pillars
Why CentraKit feels different from a starter kit
The useful distinction is simple: CentraKit is AI-native for the people building the codebase and for the customers using the product.
AI-native customer workflows
Ask, Plan, and Agent-style workflows belong in the SaaS model from day one, with tenant context and permissions already in view.
Multiple product interfaces, one codebase
Ship the SaaS app users log into, the REST API partners integrate, MCP for agent clients, CLI for power users, and docs that track the truth.
Provider-aware boundaries
Contracts in packages/data, provider code in packages/supabase, and wiring in packages/core keep humans and agents from crossing the streams.
A broader starting line than a dashboard template
The feature set is intentionally connected across product, API, docs, and developer workflow so buyers can see how the system extends.
Supabase Auth, organizations, invitations, users, teams, and role-aware access.
Tasks, quotes, customers, products, categories, brands, tags, and workflow operations.
REST API routes, OpenAPI generation, typed client output, and MCP server surface.
Agent instructions, package boundaries, tests, linting, formatting, and verification commands.
Need the implementation detail next?
Jump to the developers page for repo layout, commands, and screenshots, or go straight to pricing when you are ready to buy.