Point it. Capture it. Ship it.

Vynix turns point-and-click feedback on any live website into structured context (element, DOM, console and network) that your AI coding agent can act on instantly.

  • Zero runtime dependencies
  • Under 16KB gzipped
  • Drops into any stack
A live site with Vynix installed

0

follow-up questions for your agent

2

ways to install: script tag or extension

<16KB

gzipped, zero runtime dependencies

6+

AI agents supported over MCP

Speaks the language of your agent & your stack

Claude CodeCursorGitHub CopilotWindsurfGeminiAny MCP clientStatic HTMLPHPWordPressLaravelReactVueNext.jsSvelteClaude CodeCursorGitHub CopilotWindsurfGeminiAny MCP clientStatic HTMLPHPWordPressLaravelReactVueNext.jsSvelte
The problem

Vague feedback is expensive

“Make this bigger” sends a human (or an agent) on a guessing expedition. Vynix replaces the back-and-forth with exact, actionable context.

Without Vynix
  • “Make this button bigger.”

    Which button? Which page? What size? Cue a screenshot exchange.

  • “This looks broken on mobile.”

    No element, no viewport, no console error, just a feeling.

  • “The checkout feels off.”

    The agent guesses the selector and edits the wrong component.

With Vynix
  • Exact element: CSS selector, XPath and DOM snippet attached.
  • The page, viewport and computed styles, captured automatically.
  • Console errors and failed requests, so “broken” comes with proof.
  • A clean Markdown report, or live access over MCP.
How it works

From a click to a committed fix

Four steps. No screenshots, no selector hunting, no follow-up questions.

01

Point at anything

Click an element or drag a region on your live site. The hovered element highlights and names itself (button, link, input, image) so you always hit the right target.

02

Leave a note

Describe the change in plain language and set a type and priority. Vynix silently records the surrounding context as you write.

03

Capture the context

Page, CSS selector, XPath, DOM snippet, computed styles, viewport, plus recent console errors and failed network requests, attached automatically.

04

Ship the fix

Copy a clean Markdown report or let your agent read it live over MCP. It already knows the element, the intent, and the evidence.

Features

Everything an agent needs, nothing it doesn’t

Vynix is built for one job: turning a point-and-click into context a coding agent can act on immediately.

Two ways to capture

Click an element or drag to select a region. Each mode lands a precise, numbered pin, and the hovered element shows its name and selector so you always hit the right target.

Live element inspector

As you annotate, Vynix shows the resolved CSS selector, the React or Vue component behind the element, and its computed styles, so you see exactly what your agent will receive.

Structured context, not screenshots

CSS selector, XPath, tag, classes, inner text, dimensions and computed styles: the things an agent can actually grep and act on.

One-click GitHub issues

Turn any note (or a whole batch) into GitHub issues, and auto-assign them to Copilot or a teammate with the selector, intent and evidence already in the body.

AI diagnosis on demand

Ask Vynix to diagnose a note: it reads the captured page, element, DOM and runtime errors and returns ranked root causes, a confidence score, a suggested fix, and the files it most likely lives in.

Share without an account

Send a read-only review link to a client or stakeholder. They see every annotation in full context, no login required, and reporter emails stay private. Revoke any time.

Runtime diagnostics

Recent console errors and failed network requests ride along with each note, so “it’s broken” comes with the actual evidence.

Native MCP server

Your agent reads, clarifies, and resolves annotations live. No copy-paste, just “address my feedback” or “fix annotation 3.”

Secure by default

Optional two-factor authentication, verified email sign-up, social login, and provider keys encrypted at rest. Your account and your AI keys stay yours.

Many notes, one handoff

Drop pins across several pages, review and edit them as one numbered list, then send the whole batch (to the dashboard, a Markdown report, or GitHub) in a single click.

Three export formats

The same notes render as a human-readable Markdown report, an imperative AI prompt, or structured JSON: pick what your workflow needs from the toolbar.

Privacy by design

Diagnostics are metadata only: no request bodies, no headers, and every URL is stripped to its path. Opt out with one attribute.

Isolated by Shadow DOM

The widget renders in a Shadow DOM, so it never inherits or leaks your styles. It looks the same on every stack.

Light, dark & reduced-motion

The widget matches your site automatically and honors prefers-reduced-motion. Freeze animations to annotate a single frame.

Feather-light

Under 16KB gzipped with zero runtime dependencies. It loads with defer and never blocks your page.

What gets captured

Context that reads like a perfect bug report

Every annotation arrives with the technical detail an agent would otherwise have to ask for, assembled silently as you type.

Page

  • URL & title
  • Viewport size
  • Browser & device
  • Captured timestamp

Element

  • CSS selector
  • XPath
  • Tag, id & classes
  • Inner text
  • Box & position

Appearance

  • Computed styles
  • DOM snippet
  • Nearest landmark
  • Region screenshot bounds

Runtime

  • Console errors
  • Failed requests
  • Status codes
  • Request timing
Privacy-safe: metadata only, query strings stripped
vynix-feedback.md
## Vynix feedback: acme.store

### #1 · Headline · High
> Tighten this headline and raise contrast, fails AA.

- url: /
- selector: h1.hero__title
- xpath: /html/body/header/h1
- viewport: 1440×900
- styles: color #94a3b8; font-size 28px
- console: TypeError: cannot read 'focus' of null (hero.js:42)
claude code: vynix mcp
$ what feedback do I have on acme.store?

 3 open annotations in this batch
 #1 headline · #2 CTA button · #3 pricing card

$ fix annotation 2
 #2 button.cta: “make primary, larger tap target”
 edited Button.tsx (min-h-12, variant=primary)
 marked annotation resolved
Model Context Protocol

Your agent reads the feedback directly

Vynix ships a first-party MCP server. Connect it once and your agent can list, clarify, and resolve annotations without a single copy-paste.

  • Ask “what feedback do I have?” and get the full batch with context.
  • Say “fix annotation 3” and the agent acts on the exact element.
  • Resolve annotations in place. The dashboard stays in sync.
Claude CodeCursorGitHub CopilotWindsurfGeminiAny MCP client
Privacy first

Your context stays yours

Vynix captures only what an annotation needs, and nothing it does not.

  • No passwords collected
  • No cookies collected
  • No tokens collected
  • Screenshots encrypted
  • Private by default
FAQ

Questions, answered

A screenshot tells a human what to look at. Vynix tells an AI agent what to change: the exact CSS selector, XPath, DOM snippet, computed styles, and your intent. That is context an agent can grep and act on without guessing.

The Model Context Protocol server exposes your annotations to AI coding agents directly. Instead of copying Markdown, you can say “what feedback do I have?” or “fix annotation 3,” and the agent reads, clarifies, and resolves it in place. It can also run an AI diagnosis, read attached screenshots, and open GitHub issues over MCP.

Yes. The AI Context Engine reads a note’s captured page, element, DOM and runtime errors and returns ranked root causes, a confidence score, a suggested fix, and the source files it most likely lives in. Bring your own key for OpenAI, Anthropic, Gemini, OpenRouter, Mistral, or Groq; keys are encrypted at rest and used only for your own requests, and every run is metered so you see the cost.

Yes. Generate a read-only share link for a project and send it to a client or stakeholder. They see every annotation in full context with no login, reporter emails are hidden, and you can revoke the link or set it to expire at any time.

You can turn on two-factor authentication (TOTP) with single-use backup codes, sign in with Google, GitHub or X, and verify your email on sign-up. Passwords reset over a secure, single-use link, and sensitive secrets, including your AI provider keys, are hashed or encrypted at rest, never stored in the clear.

Yes. The widget is a single script that runs on static HTML, PHP, WordPress, Laravel, React, Vue, Next.js and more. It is fully isolated in a Shadow DOM, so it never depends on or interferes with your app.

Runtime diagnostics are metadata only. Vynix never records request or response bodies or headers, and every captured URL is reduced to its path. Query strings are stripped before anything leaves the page. You can disable diagnostics entirely with a single data attribute.

No. The widget is under 16KB gzipped, has zero runtime dependencies, and loads deferred so it never blocks rendering. Animations respect prefers-reduced-motion.

Vynix launches free while we grow. The plan is to keep the core annotation flow free and add team and automation features over time.

Stop describing bugs. Point at them.

Add one script tag and give your AI agent the context it has been missing. Free while we grow.