Getting started
Access.
Two ways to work with Aestheria, the design context layer: the free MCP your agents query over the wire, or a one-time self-host license you bring into your own project. Same system underneath. Bespoke sits alongside both when you want it shaped to your product.
Choosing the right option
Three paths, by how much you want to run yourself. MCP is the fastest start. Self-host hands you the files. Bespoke adapts the system to your product.
| Option | Best for | Billing | What you run |
|---|---|---|---|
| MCP | Agents that should build on-system from the first prompt | Free | Nothing. Query the hosted server. |
| Self-host | Teams that want the system in their own repo, versioned and owned | One-time license | The package: in your project, or on your own server, serving your team over MCP. |
| Bespoke | Teams that want Aestheria shaped to an existing product or system | By engagement | A tailored implementation. |
See plans and pricing
MCP access
MCP access is the fastest way to connect AI coding agents to Aestheria’s design context.
With this option, agents can query Aestheria’s themes, components, patterns, and usage rules through MCP. This helps tools such as Cursor, Claude Code, and other agentic development environments generate interfaces that follow the system instead of guessing from incomplete context.
Free, with every theme. No per-brand tiers: sign up, pick a starting theme, and switch among them all anytime. The same account opens these docs as your reference, with the full charts, patterns, and templates libraries.
This option is best for teams that want to start building with Aestheria quickly, without hosting the full Studio or customizing the system from day one.
Use MCP access if you want to:
- Give AI agents structured design context immediately
- Use Aestheria’s themes and patterns as provided
- Reduce off-system UI generation from AI tools
- Test how Aestheria fits into your AI development workflow
- Start lightweight before moving into a deeper implementation
What this includes:
- The hosted MCP server, every brand and component, switchable anytime
- Machine-readable guidance for tokens, components, patterns, and rules
- Full access to these docs as your design system reference, including the complete charts, patterns, and templates libraries
- Agent-readable structure for building more consistent product UI
- Support for AI code generation workflows
MCP access is ideal when your main goal is to help agents build better interfaces now.
Self-hosted source of truth
The self-hosted option is for teams that want to own the design context layer directly.
With self-hosting, Aestheria becomes your internal source of truth. Your team can host the Studio inside your own environment, customize the system to match your product language, and maintain a living design layer that both people and agents can use.
Designers, engineers, and product teams can browse the system. AI agents can query it through MCP. Everyone works from the same source instead of relying on scattered files, stale documentation, or one-off prompts.
Use self-hosted if you want to:
- Own and control the design context layer
- Customize the system for your product, brand, and workflows
- Maintain Aestheria as an internal source of truth
- Serve design context to AI agents from your own environment
- Support design, engineering, and product teams from one shared system
- Build a long-term foundation for AI-native product development
What this includes:
- A self-hosted Studio
- A browsable design system for your team
- Machine-readable design context for AI agents
- MCP-based access for agent workflows
- Customizable tokens, components, themes, patterns, and rules
- A system that can evolve with your product over time
Self-hosted is the best option for teams that want Aestheria to become part of their product infrastructure.
Bespoke
Aestheria can be used out of the box, but many teams need more than a base system.
For teams with existing products, complex workflows, or mature design systems, Aestheria can be customized to fit the way your organization already works. This may include adapting the visual language, extending components, mapping product-specific patterns, creating custom templates, or shaping the system around your AI development workflow.
A bespoke engagement helps teams turn Aestheria from a general design context layer into a product-specific source of truth.
Consulting can help with:
- Custom theme creation
- Token and component adaptation
- Existing design system migration
- Product-specific pattern definition
- AI-native workflow design
- Agent console and AI interaction pattern design
- Enterprise SaaS template customization
- MCP setup and agent workflow integration
- Design system governance
- Design, engineering, and product adoption
This is best for teams that want Aestheria to reflect their own product language, not just run the default system.
Available by engagement today. A faster, partly automated matching service follows once MCP and self-host are out.
Recommended paths
For early exploration
Start with MCP access. This gives your AI agents immediate access to Aestheria’s base design context with the lowest setup effort.
For product teams building seriously with AI codegen
Start with self-hosted. This gives your team control over the system and allows Aestheria to become the shared source of truth for both humans and agents.
For larger organizations
Self-host with a bespoke engagement. The strongest path when you want Aestheria adapted to your product, integrated into your workflow, and adopted across design, engineering, and product teams.
Summary
Aestheria can start small or become core product infrastructure.
- Use MCP when you want agents to build with better design context immediately.
- Use self-host when you want to own and customize the system as your internal source of truth.
- Use Bespoke when you want Aestheria shaped around your product, organization, and AI-native workflows.
No matter which path you choose, the goal is the same: one design context layer that people can browse, teams can maintain, and AI agents can query every time they build.