Skip to content
dijital pazarlama ajansı: clickzone
HUBTR

Ad creative performance prediction via MCP, from your IDE

VIzone Pre-Flight is an MCP server that connects directly to your AI assistant. Upload the ad creative — get a 7-day CTR, ROAS prediction, fatigue duration and concrete revision guidance based on your brand's real data, instantly.

Request AccessWatch Video
10
Quality Dimensions
7 days
Prediction Range
~8s
Analysis Time
Model Context Protocol (MCP)
IDE / AI Client
Cursor, Claude Desktop, VS Code…
MCP Protocol
Streamable HTTP transport
VIzone Pre-flight
MCP Server
Analysis Result
CTR, ROAS, revision guidance

VIzone Pre-Flight is an MCP server — the AI assistant in your IDE connects directly, setup is one line of JSON.

How It Works with MCP

An MCP server, a prediction calibrated to your brand's data

Not a generic score — a prediction based on your brand's real CTR/ROAS distribution over MCP. Without leaving your IDE.

1
MCP Connection
The AI assistant in your IDE connects to the VIzone server via the MCP protocol. It automatically pulls your brand's ad performance data (CTR, ROAS, fatigue).
2
AI Visual Analysis
Your new creative is sent to the server over MCP, scored across 10 quality dimensions and compared with your brand's winner/poor performer visuals.
3
Result Returns to the IDE
The 7-day CTR/ROAS range, fatigue duration and concrete revision guidance come straight to your IDE over MCP — no copy-paste.
Discover in 30 Seconds

How does Pre-Flight work?

Example Output

Data-driven performance prediction

A 7-day prediction, a 10-dimension quality score and a prioritized revision list.

Performance Prediction

7-Day CTR
2.1% — 3.5%
Median: 2.8%
Estimated ROAS
4.2x — 7.0x
Median: 5.6x
Fatigue Duration
8 — 18 days
Median: 12 days
Confidence Level
High
Based on 73 ads

Quality Scores (out of 10)

Visual Impact
8.5
CTA Clarity
7.2
Brand Fit
9.0
Composition
7.8
Text Legibility
5.5
Color Harmony
8.2

Revision Guidance

Critical
Move the CTA button to bottom right — increase contrast — in winner visuals the CTA is always prominent.
Important
Increase small text by 40% — readability on mobile is critical, font should be min 24px.
Important
Center the main product — shrink other elements — the visual focus is scattered.
Suggestion
Match the background color to the brand — brand consistency score is low.
MCP Server Features

Protect your ad budget

Spot potential issues before publishing the visual and set your performance expectations — all over MCP in your IDE.

7-Day Performance Prediction
CTR, ROAS and fatigue duration range — a statistical prediction based on your brand's real distribution.
10-Dimension Quality Analysis
Visual impact, CTA clarity, brand fit, text legibility, composition and more — each out of 10.
Concrete Revision Guidance
Actionable steps prioritized as critical/important/suggestion.
Winner Comparison
Your visual is compared with your brand's best and worst performing ads.
Landing Page Analysis
When an LP URL is provided, page speed, mobile friendliness and message match are also included in the evaluation.
Ad Policy Check
Risks of non-compliance with Meta and Google ad policies are detected automatically.
MCP-Compatible Platforms

Ready in every IDE that supports MCP

VIzone Pre-Flight is an MCP server — it connects to every editor that supports the protocol with 5 lines of JSON.

Cursor
Settings → MCP
Claude Desktop
claude_desktop_config.json
Claude Code
.mcp.json
VS Code
settings.json
Codex
codex_config.json
MCP Server Setup

Add it to your MCP config, done

1

Get an API Key

Request your API key for the VIzone Pre-Flight MCP server from your CLICKZONE team. A separate key is generated for each brand.

2

Add to MCP Config

Add the JSON on the right to your IDE's MCP server settings. Streamable HTTP transport — no extra setup needed.

3

Start Using It with MCP

Your AI assistant connects to the server over MCP. Say "predict this visual's performance" — the result comes straight to your IDE.

settings.json
JSON
{
  "mcpServers": {
    "clickzone-preflight": {
      "url": "https://preflight.clickzone.io/mcp",
      "headers": {
        "Authorization": "Bearer <api-key>"
      }
    }
  }
}
Frequently Asked Questions

Common questions

What is MCP?
Model Context Protocol — an open standard that lets AI assistants communicate securely with external tools. Cursor, Claude Desktop and other IDEs support this protocol.
Do my images stay on your server?
No. Images are processed only during analysis and are never stored anywhere. The whole operation happens instantly.
Which ad platforms does it support?
Meta (Facebook/Instagram) and Google Ads visuals can be analyzed. Feed, stories and reels placements are supported.
How accurate are the predictions?
Predictions are based on your brand's own ad data. The more data there is, the narrower and more accurate the prediction. A confidence level (low/medium/high) is stated with every result.
How long does setup take?
After getting an API key, adding 5 lines of JSON to the config file is enough. Ready in 2 minutes.
Is landing page analysis required?
No, it's optional. But for conversion- and traffic-focused campaigns, providing the LP URL makes the ROAS prediction more accurate.

Connect the MCP server,
and pre-flight your visual

Add the VIzone Pre-Flight MCP server to your IDE — one line of JSON, instant analysis. Contact your CLICKZONE team for an API key.

Request AccessSetup Steps
MCP-compatible: Cursor · Claude Desktop · Claude Code · VS Code · Codex