Capture a walkthrough on the page
Start a recording in Chrome and use the page as you normally would. Eyekit pairs your spoken feedback with the page elements around your pointer and the browser context needed to explain what you mean.
About Eyekit
Show your coding agent what you mean
Eyekit is a Chrome extension for people building web interfaces with coding agents. Point to part of a page, say what should change, and Eyekit pairs your spoken request with DOM context and screenshots in a session your agent can read.
What Eyekit does
Say the details that are hard to show
Talk through what you want changed as you move around the page. Eyekit transcribes your speech and includes it with the captured context. Screenshots are captured automatically for a visual reference.
Hand your agent a session it can read
Eyekit saves each session in your browser with one prompt, JSON data, and screenshots. Download a ZIP from Settings → Sessions, then give your coding agent the copied prompt and access to the ZIP or extracted local folder.
What makes Eyekit useful
Start with the change you want to make
Eyekit is built for the moment you see something in a web interface that you want to change. It follows your walkthrough and helps you explain the request to a coding agent, rather than routing it through a team bug-triage process.
Use the coding agent you already work with
There is no required in-product connection to a particular agent. You choose which agent gets the files, and it needs access to your project. Eyekit prepares context; you stay in control of the handoff and review the code changes.
Who uses Eyekit
Eyekit is for people building web applications with coding agents, especially:
- Frontend and web developers iterating on an interface
- Technical founders and solo builders working on a web product
- Small software teams using coding agents to build or debug web pages
The person behind Eyekit
I’m Uriel, a software engineer and open-source developer. I started Eyekit because I wanted a simpler way to show a coding agent what was happening in the browser, without first assembling screenshots, copied page details, and a long prompt.
For help with setup or a recording, email [email protected].
How Eyekit works
- 1
Create an Eyekit account and install the beta extension in desktop Chrome using Developer mode.
- 2
Connect the extension to your account and allow microphone access. Eyekit also needs an internet connection for transcription.
- 3
Start a capture on a supported page. Point to the relevant interface and talk through the change; screenshots are included.
- 4
Stop the capture. Eyekit saves the session and copies one prompt; files stay available in the Sessions tab.
- 5
Download the session ZIP and give your coding agent the prompt plus access to those local files; you review the proposed code changes.
Eyekit does not directly edit your code or require a specific coding-agent integration.
Key facts
- Company name
- Eyekit
- Type
- Chrome extension for spoken interface feedback and coding-agent handoff
- Founded
- Not published
- Founder
- Uriel — software engineer and open-source developer
- Headquarters
- Not published
- Website
- eyekit.dev
- Core offering
- Pairs spoken feedback with DOM context and screenshots, then exports a local session and prompt for a coding agent
- Pricing
- New accounts receive a one-time 120-minute transcription trial. Pro costs $10/month and includes unlimited transcription during the paid beta.
- Contract terms
- A one-time trial for new accounts and a monthly Pro subscription. See the Terms.
- Services
- Browser capture, hosted speech transcription through the Eyekit API and Groq, and local session export
- Communication
- [email protected]
- Notable clients
- Not published
- Customers served
- Not published
- Projects delivered
- Not published
- Social
- Not published
Frequently asked questions
What is Eyekit?
Eyekit captures spoken feedback and browser page context while you walk through a web interface. It exports a prompt and local session files that you can give to a coding agent.
Does Eyekit work with my coding agent?
Use the generated prompt and session folder with an agent that can read local files. Eyekit does not connect directly to every agent, and it does not edit your code. Check your agent's current file-access workflow before handing it a session.
Do I need a Groq API key?
No. Eyekit sends recording audio through its API to Groq for transcription, so you do not need to create or provide a Groq API key. Transcription uses your Eyekit account's trial or Pro plan.
Is Eyekit free?
New accounts receive a one-time 120-minute transcription trial. Pro costs $10 per month and includes unlimited transcription during the paid beta.
What happens to my audio and page context?
Only microphone audio is sent through the Eyekit API to Groq for transcription, alongside necessary account and usage metadata. Captured page context, screenshots, returned transcripts, notes, and voice amendments stay in this browser. Eyekit does not upload session files or send notes to a text model. Copy and Download let you choose the handoff. Screenshots may show sensitive page content; an agent you give the files to may process them under its own provider’s terms. Read the Privacy page for details.
Is Eyekit a bug tracker?
No. Eyekit makes a local session to help explain a browser issue or interface change to a coding agent. It does not replace your issue tracker or team workflow.
What do I need to use Eyekit?
You need desktop Chrome, an Eyekit account, an internet connection, microphone permission, and a coding agent that can read local files. The beta extension is installed through Chrome Developer mode. Browser-internal pages and some restricted pages cannot be recorded.