February 25, 2026
Introducing Blutui Context
The conversation around AI in agencies has become noisy. Tools everywhere. Promises everywhere. Real advantage, rare.
At Blutui, we have been focused on something more practical. How agencies actually deliver better work, faster, without bloating cost, complexity or token spend.
Today we are announcing a major enhancement to the Blutui AI development workflow, powered by a hybrid Skills deployment model and Blutui MCP packaged as Blutui Context.
This is purpose driven infrastructure for an staggeringagency advantage.
Blutui Context: A Smarter Architecture for AI Inside the Agency
We are refining how we talk about Blutui Context because precision matters.
Blutui Context is not built purely on Skills. It is a deliberate combination of:
Custom instructions
Agent skills
Blutui MCP
Each has a distinct role. Each has a different activation model. Together, they create a tightly controlled, highly efficient AI delivery system inside an Agency Web Ecosystem.
Why Not Just Agent Skills?
Because Skills and custom instructions behave differently inside an AI agent.
Skills
Activate only when required
Trigger based on the user’s prompt
Load domain specific capability on demand
Custom instructions
Are always active
Define core principles
Anchor the agent’s behaviour and output standards
If you activate everything at once, you waste context window, inflate token use and dilute precision.
Blutui Context avoids that.
The Hybrid Activation Model
Blutui Context uses a layered approach.
Core Blutui project principles are always active through custom instructions
Domain specific Skills activate only when the task requires them
For example:
Creating a form
Building a collection
Structuring dynamic content
Managing reusable components
Those domain specific capabilities are not permanently loaded. They are activated only when the agent determines they are needed to fulfil the prompt.
This ensures the agent uses its context window intelligently rather than flooding it with unnecessary knowledge.
The result is better performance, tighter outputs and far more efficient token consumption.
Efficiency is not cosmetic. In agency environments, it is commercial.
What Blutui Context Actually Is
Blutui Context is made up of three components:
1. Custom Instructions
These foundational instructions define how a Blutui project should be structured and built. They encode the principles, constraints and best practice logic that underpin every project.
2. Skills
These provide modular, task specific execution capability. They activate only when required and bring in domain depth for specific build activities.
3. MCP
Blutui MCP has two clear responsibilities:
Create, list or retrieve project resources via the API
Search Blutui documentation for additional context via the API
MCP does not replace knowledge. It enables controlled interaction with the live project and the documentation layer.
How Context Works in Practice
When combined, the system becomes powerful.
MCP can:
Create a resource inside a project
Retrieve or list existing resources
Pull in relevant documentation
Custom instructions ensure:
The resource aligns with Blutui build principles
Skills ensure:
The resource is implemented correctly for its domain use case
So the agent can create a resource via MCP and then use Blutui specific knowledge to determine how that resource should appear and function on a website.
This is orchestration, not prompting.
Why This Matters for Agencies
Legacy CMS platforms were never designed with structured AI orchestration in mind. They rely on manual workflows or loosely connected automation layers.
Blutui was built as an AWE, an Agency Web Ecosystem.
Blutui Context strengthens that position by:
Controlling AI overhead
Preventing unnecessary context inflation
Encoding agency level best practice into system logic
Allowing domain depth without permanent bloat
Turning AI from experimental tool into production infrastructure
Agencies competing against non Adland players need structural advantage, not surface level automation.
Blutui Context provides that structure.
Custom instructions for foundation.
Skills for precision.
MCP for controlled execution.
Together, they make AI commercially viable inside real agency delivery environments.