Chrome Extension · MV3 · Released 2026

Review work, reimagined as a cockpit

WireFlow is a glassmorphic workflow assistant for review pages. AI checks, font validation, and actions unified in one visual automation dock.

See in Action How It Works?
No API keys Human-in-the-loop MV3 architecture
wirestock.io/paid-projects/
Submitted
Approve
Revise
Reject
Problem Explanation
In congruent triangles ABC and MNK, angle A equals angle M, angle B equals angle N, and angle C equals angle K. Which sides are correspondingly equal?
🧩 Knowledge Check via ChatGPT
⚠
Revise

AI verdict received from ChatGPT.

REVISE: The solution incorrectly states the angle correspondences before giving the correct side correspondences.
PROPOSED CORRECT SOLUTION

Since angle A → M, angle B → N, angle C → K, the corresponding sides are: AB corresponds to MN, BC to NK, AC to MK.

0000% faster review sessions
ALL top AI platforms integrated
MV3 Chrome architecture
ZERO API keys required
Live

See WireFlow in Action

PRODUCT DEMO

Architecture

Modular by design

Each review module is a self-contained content script. The service worker coordinates. The dock is the UI shell.

background.js MV3 service worker
content/core.js Dock, control panel, settings, module injection
content/modules/kbe-parse.js Knowledge Check panel and Wirestock parsing
content/modules/kbe-paste.js Chatbot handoff and verdict watcher
content/modules/fon-check.js Font Check UI and matching engine
content/modules/wf-builder.js Rule builder and automation engine
font-check-companion.py Local PSD font parser (Python 3.8+)
workflow.compile 
// Knowledge Check loop
wireflow
  .capture(submissionContext)
  .buildGrid({
    problemImage,
    problemExplanation,
    solutionImage,
    solutionDescription
  })
  .handoff(chatbot.selected)
  .watch('RESULT_V1')
  .parse({ verdict, proposedSolution })
  .sync(afterHumanApproval);

// Font Check loop
wireflow
  .extractFonts(psd | figma | slides)
  .match(submittedFontMetadata)
  .recommend('approve' | 'revise')
  .sync(wirestockFields);

MODULARITY

Everything a reviewer needs,
nothing they don't

WireFlow injects a floating dock into Wirestock review pages. Each module handles a distinct part of the review loop.

📡 New in 2.3

Broadcast Verdict

Broadcast Verdict sends one selected outcome to every open Wirestock submission tab. Pick Approve, Revise, or Reject once, then confirm the batch from the panel. One repeated decision can cover hundreds of data sets without reopening the same modal over and over.

  • Cycles through each submission tab so backgrounded React apps mount before actions run.
  • Opens and fills each reason modal first, then confirms them in order with short waits to avoid input races.
  • Requires a two-step panel confirmation to prevent accidental mass actions.
Mass action Review control
🧩 New in 2.2

Knowledge Check

Builds a 2×2 evidence grid from the submission, sends it to the selected chatbot, reads the tagged verdict, and helps reviewers approve or revise with zero manual copy-paste.

  • Structured grid: problem image, explanation, solution image, description
  • Chatbot handoff + result parsing across Claude, ChatGPT, Gemini, Grok, Copilot
  • Revision sync with one click
AI-powered Multi-chatbot
🔠 New in 2.1

Font Check

Compares submitted font metadata against PSD, Figma, and Google Slides sources. Smart matching across family, style, weight, color, and size with approve/revise recommendation.

  • PSD parsing via local Font Companion
  • Figma and Slides extraction
  • Sync support to Wirestock fields
Font validation
⚡

WF Builder

Visual IF/THEN automation for Wirestock form behavior. Pick page elements, define conditions, fire actions, no coding required. Runs on dynamic pages with mutation observer support.

  • Field, URL, and element conditions
  • Open URL, fill field, click element, show message
  • Fire-once loop prevention
No-code Automation
✅

Approve + Next

One-click approval with session counter and automatic navigation to the next submission. Tracks your session throughput and blacklists specific IDs you want to skip.

Core
📋

Session History

Floating history panel showing recent actions with timestamps. Exports to CSV. Tracks approvals, revisions, parses, and KBE verdicts across your session.

CSV export
⚙️

Control Panel

Dock order, visibility, editable icons, timing settings, and module configuration in one place. Drag to reorder, check to show/hide any module, per-chatbot settings for AI handoff.

Customization
🧱 Design

Modular by Design

Every module is a self-contained content script: independent, swappable, and zero-dependency. Drop into a new project and you're operational on day zero, no wiring required.

  • Each module runs in isolation with no shared state
  • Enable, disable, or reorder from the Control Panel
  • New modules slot in without touching existing logic
Architecture Extensible

The Knowledge Check Loop

From submission to verdict in four steps

WireFlow treats the browser as the integration layer. No API keys. No hidden calls. The reviewer stays in control at every step.

Capture evidence

WireFlow reads the submission page: problem image, problem explanation, solution image, and solution description. Then it packages them into a structured 2×2 grid.

Compose & handoff

The grid is pasted into the selected AI assistant's chat interface (ChatGPT, Claude, Gemini, Grok, or Copilot) through the visible UI, with no API calls and no credentials.

Parse tagged verdict

WireFlow watches for the chatbot's response, extracts the APPROVE or REVISE verdict from structured tags, and surfaces the proposed correct solution directly in the panel.

Approve or sync revision

One click syncs the proposed solution back into Wirestock's revision field and clicks the Revise button, or approves the submission and moves to the next. History logged automatically.

Engineering Details

Built for speed, traceability,
and controlled automation

Manifest V3 core

Service worker plus modular content scripts. Portable, loadable, browser-native. No deprecated MV2 APIs.

SPA-aware injection

Mutation observers and reinjection checks preserve the dock across dynamic navigation and late-loading pages.

Schema-first AI outputs

Tagged result blocks make AI responses parsable instead of relying on fragile free-form prose extraction.

Human-in-the-loop

All AI handoffs run through the chatbot's visible UI. Automation stays reviewable, adjustable, and reversible.

OS-aware keybinds

Mac and Windows keyboard shortcut handling for chatbot paste and send actions. Gemini-compatible paste behavior.

Multi-chatbot support

Claude, ChatGPT, Gemini, Grok, and Copilot, each with chatbot-specific paste and verdict-detection behavior.

Local Setup

Three steps to a smarter review session

WireFlow loads as an unpacked Chrome extension with no store approval needed, no telemetry, and no account required.

1

Enable Developer Mode

Open Chrome and navigate to chrome://extensions. Toggle Developer mode on in the top-right corner.

2

Load Unpacked

Click Load unpacked and select the WireFlow project folder. The dock will appear on your next Wirestock submission page.

3

Font Companion (Optional)

PSD font analysis requires Python 3.8+ and psd-tools. Run the companion server once per session:

Terminal pip install psd-tools
python font-check-
companion.py
By Pietro Impagliazzo
Pietro Impagliazzo

I built WireFlow to improve review speed AND accuracy through automated submission context aggregation

Knowledge Check hands off to the AI you already use, parses the verdict, and syncs it back. No credentials, no API keys, no waiting. Font Check eliminates the guesswork entirely. Builder lets you define your own rules and stop doing the same thing manually every session.

Get the Keys Connect on LinkedIn →
MIT License · Chrome Extension · Manifest V3