desplega.ai Newsletter

Desplega AI: Welcome Argus to the team!

Wed, 22 Apr 2026 12:57:38 GMT

We shipped. Here's what else happened.

Argus brings qa-use to desplega.ai

We’ve built Argus for non-technical users, this is the result:

On-demand QA environments, mimicking your localhost experience when using qa-use.
A real assistant, not a chatbot. Argus understands your product, it goes deep into your business goals. The more you use it, the more it learns.
One-line CI integration. Trigger Argus through API calls, in your CI/CD, to run in-depth discoveries for new PRs. Cheaper, faster, async.

→ Generate test plans, automated tests, session recordings, improve existing ones, all in a single place.

Learning is the real bottleneck

We’ve run 20+ workshops across different companies with teams ranging from 1 to 100+ individuals. We learned three things:

Eager to improve. Everyone is looking forward to adopting more AI, they want those daily chores removed from their day-to-day.
Landscape is confusing. Too many options, too much news, a mirage of promises that don’t land. Teams end up frustrated, and without control over their solutions.
No time to waste. Their mandate to learn AI while accelerating means they rush through busy work. Experimentation lacks depth without deeply honed expertise.

→ Use Argus to leap forward, or ask about our multi-day workshops.

Time to make QA your devex powerhouse...

Or keep reading newsletters. Your call.

Sent with ❤️ by desplega.sh - unsubscribe here

desplega.ai: CLI vs MCP, a massive jump.

Fri, 30 Jan 2026 12:41:05 GMT

tl;dr: We now integrate with any coding workflow through CLI, our tailored company workshops keep growing.

Hi everyone,

If you know a team that’s still coding like it’s early 2025, we'll be happy to have a short session with their team to kick-start their efforts to upgrade their software development life cycle (slides). For those smaller teams, join our open floor conversations.

❤️‍🔥 QA where you need it the most.

Our new qa-use CLI, together with our qa-use skill, allows you to create, manage, and edit all your tests directly from your repository in localhost without any integration.

What you'll get:

• 1-minute install to full setup.
• Interact directly with desplega.ai from your dev environment
• You can track versioning for tests in your repository.
• See remotely your coding agents with our browser sessions.
• Leverage existing playwright, cypress, gherkin style tests, etc.
• Generate custom reports over your tests, and ask complex questions.

👉 You have a QA superpowers when you need it the most.

We are very excited to share what’s coming next, it will change the way you think about E2E QA. Keep tuned.

How do you know your team is shipping the right stuff?
Best,

Blog | MCP | User Guide | Releases | Unsubscribe

contact@desplega.ai

Barcelona, Spain

desplega.ai: 10x your tech team this Thursday.

Tue, 20 Jan 2026 10:57:41 GMT

tl;dr: free claude-code workshop Thursday. 50+ engineers trained. Your team's next. Join us!

Hi everyone, short-and-sweet.

❤️‍🔥 Claude-code workshops

We've 4x'd our dev velocity since Nov, 10x since Oct.

We're sharing exactly how with your team, free.

What you'll get:

• 1-minute install to full setup, see free claude plugin.
• SDLC upgrade playbook, the framework behind our gains

• AI toolbox access: Plugins, skills, agents, commands, all ready after one config line.

• Dockerized workers: Multiply claude-code instances for complex workflows with your codebase to manage the checks as code.

👉 Free event Thursday; 5pm CET / 11am EST (limited spots).

50+ engineers went through these talks. Feedback's been spectacular. Book a free 1-hour session for your team.

🏃🏽‍♂️‍➡️ Alpha: Multi-layer Test Cache

Caching is now enabled by default for select tests, this means 30% faster suite execution.

What this means for you:

• Less latency on chained multi-step tests
• Fewer flaky results; faster, more reliable runs
• Full rollout February.

10k+ tests runs, and 1000+ tests created since our last update.

What's your team shipping?
Best,

Blog | MCP | User Guide | Releases | Unsubscribe

contact@desplega.ai

Barcelona, Spain

2026: bearing 3 gifts from desplega.ai

Fri, 02 Jan 2026 12:28:52 GMT

tl;dr: API Synthetic checks, 10x the value you get from your claude-code subscription, and an old school sim game.

Hi everyone,

In the spirit of these days, we bring you 3 gifts to help you start the year at full steam.

❤️‍🔥 API Synthetic Checks

Simple, fast, set-&-forget checks to make sure your API works all-the-time.

You can set them up in our UX or directly in your codebase, and have them run constantly to ensure everything works flawlessly in your production APIs.

What this means for you:

• Configure API health-checks with assertions, variables, dependencies and more.
• Receive alerts when an API starts falling, and add hooks to execute tests upon changes.

• Get issues generated whenever a check fails.

• Integrate with your codebase to manage the checks as code.

What's next:

Flakiness analysis & custom actions upon failure.

🍯 Agent Swarm: FOSS claude-code orchestrator

We decided to open our claude-code agent orchestrator after we got a 10x value from our Claude subscriptions. We already have 22 stars, and we keep pushing.

We want you to get the most of your claude-code subscriptions, It’s free and, more importantly, open source. Use it as inspiration to customize your own LLMs setup, and make sure to pick fun names for your agents. Check our videos [1][2]

We'll be happy to talk with your team on leveraging these advanced setups.

What this means for you:

• See a template for an advanced Claude set-up relying on skills, commands, agents, plugins, etc.
• Learn how to dockerize & multiply your workers, so you can have complex workflows.
• Interact with a single Claude Code agent, instead of multiple, independent, ones.
• We are happy to jump into a workshop to show your team how you could benefit from it.

What's next:

We have internal integrations with slack, and qa-use-mcp (to see what they actually did), and many other tools, we’ll be opening them as we progress.

🃏 SimQA: It’s like SimCity 3000 with more bugs

If you grew up in the 90s, you know open-ended simulation games were a thing. They had those amazingly frustrating random events, like a CEO sudden pivot or an amazing intern that changes everything… or you could be a victim of your own success with an unexpected flood of visitors from hacker news.

Check out how good or bad you are at defining a QA strategy trying to balance your immediate needs, long-term vision, and customer satisfaction.

Ping me if you get an S classification!

In less than 6 months, we have created 2.2k & ran 73k e2e tests, amounting to 560k test steps runs. We continue to aim to make your QA life more enjoyable.
Best,

Blog | MCP | User Guide | Releases | Unsubscribe

contact@desplega.ai

Barcelona, Spain

2 weeks: Lovable desplega.ai

Tue, 02 Dec 2025 12:32:50 GMT

tl;dr: New Analytics, auto-discovery, and immersive vibecoder UX!

Hi everyone,

The past 2 weeks were intense, we used this time to make desplega.ai friendlier for enterprise and vibecoders alike.

❤️‍🔥 Auto-discovery, Tinder meets QA

Inspired by conversations with our users, we decided to give a swipe flavor to auto-discoveries.

Every week, you’ll receive a set of new tests to review, and decide what to do next. We define the discovery focus based on context, personas and current coverage you have in our platform. Afterwards, you’ll be notified with a short email, and you can decide which tests to incorporate or not to your suite.

What this means for you:

• You can enable auto-discoveries in your discovery page, or click here.

• Every week between 3 to 15 new tests, specifically crafted for your app.

• Decide in minutes which ones to accept/reject.

• Review issues, and edit the generated tests to cover your exact needs.

What's next:

We will add configuration options so you can predefine the amount and frequency for these auto-discoveries.

🤓 QA Pulse: Analytics

We aim to answer your questions on Coverage, Flakiness, and help you understand where you are at as a team.

Please give us suggestions, as we consider this a key tool for those teams trying to understand their investments in QA.

What this means for you:

• Flakiness: See which tests are flakier, and why.
• Speed: diagnose why test suites or tests are slow, detect performance issues and poor flows.
• Coverage: are your tests stale, never run, or simply cancelled too often
• Pass Rate: are you a top 1% team?

What's next:

Integrations with your CI/CD & ATS will allow us to track here DER/CFR/MRR and other key metrics.

🧙 Vibecoder QA

We are getting closer to a full self-serve vibecoder experience, starting this week with an immersive Lovable.dev experience.

We have launched a chrome extension that adds a '⚗️ QA' tab in your vibecoding UX. With this, you can start automating your regression tests in seconds.

What this means for you:

• Fully automated regression tests for vibecoded apps.

• Feedback directly in your tab, copy+paste issues to your favorite LLM.

• Iterate with confidence, you know your solution works as intended.

• Run scheduled tests, remotely, even in production.
• Once your project graduates, your tests go with you.

What's next:

Replit app store & business-use integration.

As usual, if you have pressing needs, reach out through slack, email, or discord,
Best,

Blog | Chrome Extension | User Guide | Releases | Unsubscribe

contact@desplega.ai

Barcelona, Spain

This week: Auto Discovery & MCP++ ❤️‍🔥

Tue, 18 Nov 2025 15:48:21 GMT

tl;dr: You'll now have discovery running on autopilot, once a week!

Hi everyone,

We are excited to share a few updates that are meant to change the way you interact with desplega.ai in your day-to-day quality life.

🔍 Auto-discovery once a week in your inbox

Next Monday you are going to receive automatic test suggestions in your inbox, and it's going to happen weekly afterwards. Select the ones that you care about, and we'll implement them in seconds.

You'll be able to increase coverage without effort, while our system looks into all the context, personas and existing tests to generate new ones for you.

What this means for you:

• Every Monday you'll receive an email with new tests discovered

• You can review those new tests in 5 minutes, suggest changes, or keep them as is.

• You can give a thumbs up or down, and tag them so they run in your test suites.

What's next:

We will add configuration options so you can predefine the amount and frequency for these auto-discoveries.

⚡ MCP++

We have received feedback about our MCP's latency outside the EU. Starting this week, we have a server in the US – make sure your setup is configured correctly. You only need to set your MCP config var QA_USE_REGION to US.

On top of that, we have upgraded our local-tunnel to be more reliable [1][2] and we are extending the APIs and tools you can access directly or through our MCP.

What this means for you:

• Use the actual MCP server that's closer located to you

• We have now APIs for most actions in our platform, and you can try them directly from our platform.

What's next:

Integrations with your CI/CD to actually evolve the tests together with your code, reach out to us if you are interested in this.

As usual, if you have pressing needs reach out through slack, email, or discord,

contact@desplega.ai

Barcelona, Spain

This week: Agents & Back-end testing ❤️‍🔥

Mon, 20 Oct 2025 15:58:24 GMT

tl;dr: If you are thinking about how to test/monitor your back-end and front-end continuously, check out our alpha backend solution!

Hi everyone,

This past week was intense at desplega.ai. After a few conversations and many long coding sessions, we have 2 solutions in alpha that we would love to have your feedback on.

Real end-to-end: Back-end joins the party.

Verifying your business logic works as intended goes well beyond UI e2e tests. In fact, one common challenge is to ensure those simple back-end invariants happen; e.g. "When I receive a webhook from A, I trigger a call Z in service Y, and that results in F".

With that in mind we are planning on providing you with an open source library for live monitoring and comprehensive testing of your solutions, including business logic. Check this short video.

What this means for you:

• Test your backend flows, ensuring the right call stacks happen in time and manner with invariants if you wish for.

• Have e2e tests triggered before, during or after specific flows, that way you can 'check the display' once you know the data is there. This helps you with complex flows with batch processing, long provider calls, etc.

• API calls as first-class citizens on our platform. You can now test scenarios where there's no/limited UI present.

• Monitor your flows in production, receiving errors immediately upon assertion issues. For example, you'll be notified if a service call never executes even though it should have happened after a certain API call, or a business logic invariant (like a $ amount check) does not hold in a specific flow.

What's next:

We want to make this project open source, we need your feedback!

Stealth Browser Agents

In the past months we have diverged from other browser agentic solutions given key limitations with performance and reliability. We have decided to open the engine, so you can have fun finding interesting use cases: agents.desplega.ai.

What this means for you:

• Define parametric routines that you wish to execute repeatedly on web

• Stealth mode as an option for our browser agents, this helps with CAPTCHA, firewalls, and other UX blockers.

As usual, if you have pressing needs reach out through slack, email, or discord,

contact@desplega.ai

Barcelona, Spain

This week: Node-based editor support & Detailed test descriptions 🔍

Mon, 13 Oct 2025 18:43:20 GMT

tl;dr: Our agents can now interact with your workflow builder natively, and we show you the intent behind your tests.

Hi everyone,

This past week we deep dived into our platform usability.

Node base editor support

We have added native support for node-base editors implemented with React Flow. This means you can now define test cases that interact with flow builders, and evaluate if your editors are functioning as expected for all your users.

What this means for you:

• Our discovery, new test, and in general all our tests now support interacting with your flow editors

• You can add context to our platform so our agents can navigate these flow builders smartly

• You can use the full power of our platform in these workflow builders: AI actions, assertions, file upload, data extraction, etc, as with any other page

What's next:

We are evaluating our next integration in this area, reach out with suggestions!

Understanding what you are testing

We have revamped our agents to analyse and express in an Arrange-Act-Assert format each test. This is key, given that it represents our system understanding for the test intent.

What's next:

• Edit the description to re-generate your test

• Kick off a new test from that description

Other key improvements

Clipboard support – Access and control the browser clipboard as a user would do in tests

Improved agent accuracy – AI based steps and next step assessment improved

Browser configuration – Control user agent, location and localization per application, testing how you app works in different languages never got easier

As usual, if you have pressing needs reach out through slack, email, or discord,

contact@desplega.ai

Barcelona, Spain

This week: MCP & Local Storage for runs 🔍

Mon, 06 Oct 2025 11:22:14 GMT

Tl;dr: You can now see Local Storage, Cookies, Clipboard and more, also our MCP is a go.

Hi everyone,

This past week was all about accelerating developers at desplega.ai.

Desplega MCP in beta

With the added support of test promotion, and multi-app, you can now use the MCP in combination with your CI/CD pipeline.

What this means for you:

You can create new tests automatically while generating new features,
You can debug those tests locally or in our app,
You can later on decide which of those tests you want to run continuously in production, staging or kicked-off by your deployment process with a single click.
What's next:
- We will strive to make the experience more seamless between your localhost and our cloud solution.

Deeper understanding of your headless run

We had multiple conversations with several technical leads that were asking for: LocalStorage, Cookies, Clipboard actions, and access to the indexDB. All that paired with a better UX for your browser logs.

What this means for you:

🍪 Understand your cookies
You can deep-dive into the user cookies, see what's loading and potentially replicate the state locally, if you prefer to use our MCP.
🔍 Local Storage
Explore inconsistencies in your local storage, see your saved configurations or other user preferences.
🗄️ Clipboard
Coping & pasting is not going away any time soon, and ensuring you are copying the right thing could be key for some of your user flows.

We will be happy to demo this for your team, reach out!

Best,

desplega.ai
unsubscribe

This week in desplega.ai: MCP & custom browser operator 🚀🔥

Mon, 29 Sep 2025 18:54:10 GMT

tl;dr: Check out this short video for a detailed view on how to use our new MCP.

Hi everyone,

We continue to focus on acceleration at desplega.ai, and last week wasn’t the exception.

Desplega MCP in alpha

During the past few days we have been fishfooding our MCP version of deplega.ai. This is a completely new approach that will allow your dev team to work in lockstep with our solution. It basically enables any LLM CLI set up to create, run, and utilize the feedback from desplega.ai tests in real time. The outputs help you iterate on your code live.

Check the full demo video

What this means for you

If you use cursor, claude-code, codex, gemini-cli, etc, use this one-liner.

You can now interact with desplega.ai in your local dev environment
No need to wait for your CI/CD, sandbox or production environment.
Immediate feedback while you code
The output can be used to re-write your software, live.
Deeper integration
You can leverage your LLM context to generate much more precise intents on the fly, without actually integrating with us.

What’s next

Multi-app support
Soon you’ll be able to execute multiple apps, for different personas – and credentials – running more complex scenarios with inter-app dependencies.
Test promotion
With this new change, we are adding a way for you to decide when a test, or new test version, needs to be promoted to production, sandbox or other test environments.

Our very own browser operator

The first iterations of desplega.ai relied heavily on different browser operators to generate your tests. This seemed as the most cost-effective alternative at the time. Alas! We even collaborated with the most famous open-source browser operators out there, fixing issues that they themselves weren’t able to overcome.

However, in the past month it became clear to us that to provide you all with faster, reliable and accurate discovery and test runs we needed something that was beyond the state-of-the-art. During the past weeks we have been silently testing a new version of our software, our very own browser operator bundled with a specific set of agents meant to provide you with all the testing functionality you need.

What this means for you

🏃 Faster AI actions
Your complex tests, those that depend on AI for assertions and custom actions, now can run much faster due to us leveraging the specificity of our problem space.
🔍 Accurate discovery
Our discoveries are now much more accurate, instead of trying to navigate any issue to success, they are laser-focus on evaluating the quality of your user experience.
🧑‍🏫 Human intervention at any point
Now you can pause any discovery or AI action while generating tests and give clarifying tips to our agents. You will be prompted through notifications if our agents evaluate they need user input for specific tasks, so you can have an increased context when creating new tests.

Let us know if you want to be part of our MCP alpha, we will be happy to facilitate this for your team,

Best,
PS: If you are a hobbyist vibe-coder, please reach out to us! We have an early version that works with Replit, V0, Lovable, bolt.new, and more.

Ez.-
Co-founder
desplega.ai

unsubscribe

This week's update: Agent Swarms & Network Logs

Mon, 22 Sep 2025 23:04:05 GMT

Tl;dr: Check out this short video deep-diving into how to take advantage of these new features.

Hi everyone,

Last week at desplega.ai was all about accelerating your experience.

Discovery with Agent Swarms

We have simplified our test creation options, making our test discoveries faster, and more detailed. We now allow you to spawn multiple agents to discover new test cases, immediately generating automated tests.

What this means for you:

More discoveries: With this update, you can run discoveries frequently, and tag selected tests to add them to your test suites.
More control: Our agents will prompt you for key input for those key decisions. Similar to the YOLO mode we have provided in our test creation option.
Less wait, less clicks: every flow we find will be automated by default, that means now all your tests are ready to be executed and fine tuned as soon as discovery finishes.

Detailed Network Logs for Test Runs

One thing we always hear is what happens after we find a bug, and for that we created issues.

Our issues section gives you fix suggestions, has before and after screenshots, contains a video link to the test, and it will self resolve issues whenever we detect the tests now pass. However something was missing: the debugging info for your Tech team.

In that light, we have extended our network, logs and results sections to provide you with more detailed root cause information. Simply share the issue or test run directly with your Tech team, and they’ll be able to debug the issue in an instant.

Until next week,

Best,

---
Ez.-
Co-founder
desplega.ai

unsubscribe

10x Faster, and a Tool Built for You

Mon, 15 Sep 2025 11:39:21 GMT

Tl;dr: tester.desplega.ai is a completely free test-management app that you should try.

Hi everyone,

As founders building in the AI and dev tool space, we are obsessed with two things: the relentless speed of innovation and the real-world problems developers face every day. It’s easy to get lost in the hype, but true progress comes from solving practical challenges that slow teams down.

Today, I’m excited to share two major updates from desplega.ai that do just that.

A 10x Leap in Testing Velocity

We are laser focused on making it easy to create and run tests. To that end, during August we re-architected our core AI engine from the ground up, delivering a 10x improvement in both test creation and runtime speed.

This isn't just an incremental update; it's a fundamental shift. It is the result of major advancements in how our models interpret application behavior and generate resilient automation.

What this means:

Faster Feedback: Get comprehensive test results in seconds, and create them in under a minute.
Increased Confidence: Run more thorough suites without sacrificing release cadence.
Higher Efficiency: Free up your engineering team to focus on building, not waiting.

Introducing tester.desplega.ai: A Free Test Management Platform

In call after call with engineering, Ops and QA leaders, we’ve heard the same story: managing test suites, tracking results, and maintaining a source of truth is a universal pain point. It’s a tax on productivity, and it shouldn’t be. They end up with tens of tabs open, multiple different spreadsheets and configurations they need to follow for each project.

That's why today, we're releasing tester.desplega.ai. A powerful, intuitive test management platform; completely free.

We believe essential tools should be accessible to everyone. This platform is our answer to that need, built to help teams of all sizes organize, execute, and analyze their testing efforts without the overhead of another expensive tool.

Check out tester.desplega.ai and get started for free

Our mission is to build tools that solve deep technical challenges and empower the dev community. These updates are a direct reflection of that mission.

We’re excited to see what you build with them.

Best,

P.S. The new tester.desplega.ai platform has a special origin story. Just last week, we held our first-ever hackathon in Gavà. The push for this free tool was born from that intense burst of creativity and collaboration. We believe in building what the community needs, and sometimes the fastest way to do that is to just get in a terrace and build it.

---

Ez.-
Co-founder
desplega.ai

unsubscribe