EarlyAI

You code, let Early take care of the test

5.0•1 review•

324 followers

You code, let Early take care of the test

5.0•1 review•

324 followers

Visit website

Automation tools

•

AI Coding Agents

•

AI Infrastructure Tools

EarlyAI is devTool AI Agent that automatically generates high-quality, working unit tests around the application code, helping developers and companies reduce the cost of bugs and deliver higher-quality software, faster.

Free Options

Launch tags:Software Engineering•Developer Tools•Artificial Intelligence

Launch Team / Built With

Framer AI AgentsDesign and publish professional sites with AI

Promoted

EarlyAI

Maker

📌

👋 Hi Product Hunt! I'm Sharon Barr, co-founder and CEO of Early, and I’m excited to introduce you to EarlyAI, your AI-Agent test Engineer. Why write tests? Let EarlyAI do it for you, boost your code quality and do what you love most – build innovative applications fast. Special offer for our voters today, 6 months free and early access to early releases of Early. Our Solution EarlyAI is an intelligent DevTool that enhances code quality by generating high-coverage, working unit tests: • Green Unit Tests: Protect your code from bugs. • Red Unit Tests: Identify potential issues in your code. • Generating high coverage for your code in a click • Plus, EarlyAI suggests documentation and improvements to your methods. Our vision is to liberate developers from bugs, helping you build high-quality software, faster. Key Features & Capabilities: 🪄Auto Test Generation: Generate high quality unit tests for your methods in a click. 👀 Easy navigation the new unit tests: Access newly generated unit tests via EarlyAI View, CodeLens, Context Window, or through Earl (see below). 🔧 Documentation and improvement suggestions: Enhance code readability with auto-suggested documentation. 🚀 Pull Request testing in a click: Easily generate high-coverage tests for your pull requests with a dedicated “My Changes” view. ✅ High-Quality Tests & Increased Coverage: Generate 30 unit tests in a minute or 1,000 unit tests under an hour for your project. Check our blog to see how we achieved 70% coverage with over 1,000 unit tests and 13,000 lines of test code on the ts-morph OSS project. Meet Earl, your AI Test Engineer When tests are running, Earl steps in to take care of your quality engineering needs: 1. Test Generation in Progress: See the status of test generation instantly 2. Simple navigation: Quickly locate methods and corresponding tests 3. Method Documentation: Insert relevant documentation to improve maintainability 4. Code Improvement Suggestions: Receive actionable insights for code enhancement 5. More capabilities are coming soon! What Users Have Achieved with EarlyAI With EarlyAI, our users have already generated over 50,000 unit tests, increased code coverage, identified bugs, reduced the cost of bugs, and accelerated development cycles. We invite you to try Early! Search for EarlyAI on VSCode Marketplace, install, and generate your first tests. Your feedback is invaluable as we continue to enhance the developer experience. You code, let Early take care of the tests. Special thanks to Ben Lang @benln for hunting us today and helping us bring EarlyAI to Product Hunt! Yours, Sharon Barr CEO & Co-Founder of EarlyAI startearly.ai

Report

2yr ago

Databar.ai

@benln @sharon_barr2 I love how Earl makes the whole process more intuitive and user-friendly. The fact that you've already helped developers generate over 50,000 unit tests shows there's a real need for this! Congrats on the launch!

Report

2yr ago

@benln @sharon_barr2 Hey Sharon, Congrats on the launch. The Other day i was looking a similar niche product Owlity AI. Do you see a good differentiation between what you offer now. A quick glimpse will help me decide better.

Report

2yr ago

EarlyAI

Maker

@benln @imraju thanks for bringing this up. i was not aware of Owlity, and from what i see on their website, they are servicing different purpose and different personas/users compared to Early. Early is a devtool for developers and not for QA personal (maybe QE ones). Early come to take the effort of writing unit tests which is part of the code development task during (and not after) the development. If you get high quality unit tests before you submit your PR you caught the bugs at the earliest time possible. Catching a bug on any later step will cost more time, money, and work. Check this short summary of a webinar we did with the core contributor of ExpressoTS and how he used Early:

Report

2yr ago

Vela Partners

Congrats! Do you have any performance metrics published? I'd be curious to know the performance on next.js as UI is the place where we spend most of our time testing.

Report

2yr ago

EarlyAI

Maker

@yigit depends what you mean by performance. in our world, performance means how good the tests we generate are, are they really helping you increasing the quality of your code by finding and fixing bugs, understanding your code better, etc. We measure the quality of the tests themselves by running mutation tools on the results. you can read more on this blog: https://www.startearly.ai/post/o... happy to have a chat if you'd like to learn more.

Report

2yr ago

Vela Partners

@sharon_barr2 I'd mostly be interested in UI-focused testing performance and ROI on it. Does it reduce bugs being reported? Does it decrease code regressions? Any automated tests through looking into the logs or streams of how people use the product (besides the code understanding)?

Report

2yr ago

Parsagon

Very cool product! I have two questions. 1) I'm curious if you have any data on what percentage of generated tests are accepted by users vs what percentage require edits or deletion. (Used to be a program synthesis researcher, so I'm always curious about data on this kind of stuff :) ). 2) For those of us that practice test-driven development (or at least try to), do you envision EarlyAI being able to generate tests off of specs rather than code at any point? Anyway, congrats on the launch!

Report

2yr ago

EarlyAI

Maker

@sand1929 good questions 1. i don't have the data, but i can share what a Richard Zampieri discussed on our webinar today (i will post the video). he generated 206 tests (about 9,000 lines of code) that covered 89% of his code. he said that about 10% of the tests were red, and that most of them actually discovered issues in his code. he said that he did a work of a month is 1-2 days. so i guess he accepted most of the tests. as a second data point, you can check my blog here: https://www.startearly.ai/post/o... You can see that for 64% of the methods (out of 210 methods) Early generated 100% coverage. i would use that as an acceptance criteria because the mutation score for these methods is 90%, which is very high. for your second question, read this blog: https://www.startearly.ai/post/t... but in short, here is how EarlyAI can change the way TDD is done: 1. Create method signiture and document it on the comment above. 2. Generate tests using EarlyAI - mostly failing happy path tests will be generated. 3. Implement the method, tests will become green 4. Re-generate tests, more tests will show up, this time with edge-cases 5. Finish implementing the method, make sure all tests became green. Try it out and let me know what you think.

Report

2yr ago

Parsagon

@sharon_barr2 Will do! Thanks for the detailed response!

Report

2yr ago

Shift-left at its finest.. and making the robots do the heavy lifting. Congrats on the launch. It's great to see the positive use-cases for AI rolling out startup-after-startup.. what's your vision past unit-testing?

Report

2yr ago

EarlyAI

Maker

@baat_enosh Thanks. unit-testing is the the begining towards our quality for a fully autonomous AI test engineer that twill take care over all your testing needs, and more importantly, will help developers delivery high quality software, faster. You code, Early will take care of the test

Report

2yr ago

😹 LOL

I started using it and after setting it up it's so smooth! Basically, you get a whole test suite out of a component with one click!

Report

2yr ago

EarlyAI

Maker

@lorenzo_zarantonello Thank you! One interesting usecase a user told us about is that our tests forced him to look harder at his code and he found 9 issues and fixe them. Delivering higher quality code is our goal, tests are just the means to that end.

Report

2yr ago

EarlyAI

Maker

Last blog for today. Introducing EQS – Early Quality Score, redefining test quality beyond code coverage. Are your tests truly protecting you from bugs? Code coverage is often the primary metric used to assess test quality, but we believe there’s more to it. Learn how we at Early set a higher standard for evaluating “how good your tests really are.” We use the EQS standard to asses our own cutting edge technology, demonstrating how comprehensive unit tests can test every case in your code like never before. https://www.startearly.ai/post/i... If you want to learn more, ping me.

Report

2yr ago

Plano

Congrats on the launch - building test cases (or complete test coverage) while important as a engineer to think about all the edge cases, if something makes it fast/better then I am in.

Report

2yr ago

EarlyAI

Maker

@salman_paracha you are right. happy path tests are normally covered, either by unit test or all the way to e2e. the most expensive bugs, see CrowdStrike as an example, are the edge cases, out of bound array exception. and testing those at e2e is very difficult. testing it at the unit test if possible could be the best ROI.

Report

2yr ago

1 2 3

5.0

Based on 1 review

Review EarlyAI?

Reviews

Most Informative