[{"data":1,"prerenderedAt":49},["ShallowReactive",2],{"article-en-upp-natural-language-iot-testing":3},{"article":4,"related":18,"prevArticle":38,"nextArticle":42,"availableLocales":46},{"id":5,"slug":5,"title":6,"excerpt":7,"content":8,"image":9,"date":10,"published":11,"status":12,"category":13,"keywords":14,"meta_description":15,"og_image":16,"custom_css":16,"created_at":10,"updated_at":17},"upp-natural-language-iot-testing","When software testing meets the real world","Once a system starts interacting with the real world, verifying screens and APIs is no longer enough.\nThis article explains why Qaptly opened up UPP (Unified Pilot Protocol),\nand how combining it with natural language lets IoT tests verify whether things actually happened.","\u003Cp>Today we look at how to use Qaptly to test scenarios where software and hardware are integrated\u003C/p>\u003Cp>In most automated software testing tools, it is clear where \"testing\" happens:\u003C/p>\u003Cp>on the screen, in the API, inside the system.\u003C/p>\u003Cp>\u003C/p>\u003Cp>As long as all of these return the right results,\u003C/p>\u003Cp>the test is considered passed.\u003C/p>\u003Cp>\u003C/p>\u003Cp>But once a system starts interacting with the real world, things are not so simple.\u003C/p>\u003Cp>\u003C/p>\u003Ch2>\u003Cstrong>A small but very common IoT problem\u003C/strong>\u003C/h2>\u003Cp>Say you have an IoT device that can be controlled from an app or a web interface.\u003C/p>\u003Cp>There is a button on the screen: \u003Cstrong>\"Turn on light\"\u003C/strong>\u003C/p>\u003Cp>After you tap it:\u003C/p>\u003Cul>\u003Cli>\u003Cp>The UI shows \"On\"\u003C/p>\u003C/li>\u003Cli>\u003Cp>The API returns success\u003C/p>\u003C/li>\u003Cli>\u003Cp>The system state shows ON\u003C/p>\u003C/li>\u003C/ul>\u003Cp>Here is the question:&nbsp;&nbsp;\u003Cstrong>did the light actually come on?\u003C/strong>\u003C/p>\u003Cp>For a human tester, this is easy to check.\u003C/p>\u003Cp>For a testing tool, it is awkward.\u003C/p>\u003Cp>\u003C/p>\u003Ch2>\u003Cstrong>Where traditional test automation gets stuck\u003C/strong>\u003C/h2>\u003Cp>For traditional tools, this test flow usually only gets this far:\u003C/p>\u003Cul>\u003Cli>\u003Cp>Click the button\u003C/p>\u003C/li>\u003Cli>\u003Cp>Verify the screen or the API response\u003C/p>\u003C/li>\u003C/ul>\u003Cp>What it can verify is:\u003C/p>\u003Cblockquote>\u003Cp>\u003Cstrong>\"What the system says it did\"\u003C/strong>\u003C/p>\u003C/blockquote>\u003Cp>What it cannot answer is:\u003C/p>\u003Cblockquote>\u003Cp>\u003Cstrong>\"What actually happened in the real world\"\u003C/strong>\u003C/p>\u003C/blockquote>\u003Cp>If the light did not come on but the system thinks it did,\u003C/p>\u003Cp>the test still shows as passed.\u003C/p>\u003Cp>\u003C/p>\u003Ch2>\u003Cstrong>Qaptly could already control many platforms from the start\u003C/strong>\u003C/h2>\u003Cp>Before talking about UPP, there is some background to make clear.\u003C/p>\u003Cp>From the beginning, Qaptly has supported several Pilots:\u003C/p>\u003Cul>\u003Cli>\u003Cp>Browser\u003C/p>\u003C/li>\u003Cli>\u003Cp>Android\u003C/p>\u003C/li>\u003Cli>\u003Cp>iOS\u003C/p>\u003C/li>\u003Cli>\u003Cp>macOS\u003C/p>\u003C/li>\u003Cli>\u003Cp>Windows\u003C/p>\u003C/li>\u003C/ul>\u003Cp>In other words, at the level of \"operating system and screen\", Qaptly could already automate everything.\u003C/p>\u003Cp>\u003Cstrong>But the problem is not at that level.\u003C/strong>\u003C/p>\u003Cp>\u003C/p>\u003Ch2>\u003Cstrong>The problem: this way of dividing things never reaches the real world\u003C/strong>\u003C/h2>\u003Cp>However complete your platform Pilots are, you still run into the same problem:\u003C/p>\u003Cblockquote>\u003Cp>\u003Cstrong>The state of the real world does not belong to any operating system.\u003C/strong>\u003C/p>\u003C/blockquote>\u003Cp>Whether an IoT device really carried out a command, whether a sensor really reported a value, whether the hardware really entered a given state:\u003C/p>\u003Cp>none of these can be forced into being:\u003C/p>\u003Cul>\u003Cli>\u003Cp>an Android feature\u003C/p>\u003C/li>\u003Cli>\u003Cp>a Browser behaviour\u003C/p>\u003C/li>\u003Cli>\u003Cp>or the result of some UI\u003C/p>\u003C/li>\u003C/ul>\u003Cp>If a test can only hang off a \"platform Pilot\",\u003C/p>\u003Cp>it can only ever verify half the story.\u003C/p>\u003Cp>\u003C/p>\u003Ch2>\u003Cstrong>That is why UPP exists\u003C/strong>\u003C/h2>\u003Cp>The starting point for UPP (Unified Pilot Protocol) is actually simple.\u003C/p>\u003Cp>Instead of trying to cram all the hardware logic into Qaptly, we accepted one thing:\u003C/p>\u003Cblockquote>\u003Cp>\u003Cstrong>Some states can only be confirmed by \"a role that lives in that world\".\u003C/strong>\u003C/p>\u003C/blockquote>\u003Cp>UPP was originally only for our project-based clients, since these scenarios need a lot of customisation.\u003C/p>\u003Cp>With UPP, you can create a Pilot that represents an IoT device or sensor:\u003C/p>\u003Cul>\u003Cli>\u003Cp>It knows the device's real current state\u003C/p>\u003C/li>\u003Cli>\u003Cp>It can report whether the light really came on\u003C/p>\u003C/li>\u003Cli>\u003Cp>It does not need to be part of any operating system\u003C/p>\u003C/li>\u003C/ul>\u003Cp>Qaptly does not need to know how the hardware is implemented. It only needs to let the Pilot that actually controls and understands the situation do the work.\u003C/p>\u003Cp>\u003C/p>\u003Ch2>\u003Cstrong>This is where natural language matters\u003C/strong>\u003C/h2>\u003Cp>If UPP were just one more hardware verification mechanism, it would only be another engineering integration.\u003C/p>\u003Cp>But Qaptly's core has never been just \"connecting one more thing\".\u003C/p>\u003Cp>We let the test flow \u003Cstrong>exist first in natural language\u003C/strong>.\u003C/p>\u003Cp>For example:\u003C/p>\u003Cblockquote>\u003Cp>\"After I tap 'Turn on light' in the app, the device should light up within 2 seconds, and the sensor must report that the light is on.\"\u003C/p>\u003C/blockquote>\u003Cp>This is not a specification document,\u003C/p>\u003Cp>and it is not a script.\u003C/p>\u003Cp>It is how a person would naturally describe the flow.\u003C/p>\u003Cp>In this flow:\u003C/p>\u003Cul>\u003Cli>\u003Cp>Qaptly understands the conditions described in natural language\u003C/p>\u003C/li>\u003Cli>\u003Cp>The Platform Pilot operates the app or web interface\u003C/p>\u003C/li>\u003Cli>\u003Cp>The IoT Pilot reports the actual hardware state through UPP\u003C/p>\u003C/li>\u003C/ul>\u003Cp>Whether the test passes\u003C/p>\u003Cp>no longer depends only on the screen or the API,\u003C/p>\u003Cp>but on whether the thing described actually happened.\u003C/p>\u003Cp>\u003C/p>\u003Ch2>\u003Cstrong>Why we chose to open up UPP\u003C/strong>\u003C/h2>\u003Cp>Because this need will not show up in just one product.\u003C/p>\u003Cp>Any system that starts touching the real world will run into this:\u003C/p>\u003Cul>\u003Cli>\u003Cp>platform control works\u003C/p>\u003C/li>\u003Cli>\u003Cp>but state verification does not\u003C/p>\u003C/li>\u003C/ul>\u003Cp>UPP is not there to show off capabilities. It is there so the testing model itself can be extended.\u003C/p>\u003Cp>\u003C/p>\u003Cblockquote>\u003Cp>\u003Cstrong>Qaptly understands how people describe the world, and Pilots tell us whether the world works that way.\u003C/strong>\u003C/p>\u003C/blockquote>\u003Cp>We chose to open UPP up rather than keep it hidden inside the system.\u003C/p>","https://cdn.sqa.tw/blog_pictures/upp_cleanup.jpg","2026-02-01T04:30:00+00:00",true,"published","sqa","UPP,  Qaptly, natural language testing, IoT testing, IoT automation testing, end-to-end IoT testing,hardware-software integration testing","Qaptly uses natural language testing and UPP (Unified Pilot Protocol) to close the key gap in IoT\nand real-world integration testing: the system state is correct, but the physical action never happened.",null,"2026-02-03T13:33:23.295+00:00",[19,26,32],{"id":20,"slug":20,"title":21,"excerpt":22,"image":23,"date":24,"category":25},"ai-development-sdlc-release-verification","When AI builds it in two hours but verification still takes three days","Once AI speeds up development, spec sign-off, testing and release do not automatically speed up with it. Using a real case from a booking system SaaS, we describe how we adjusted the SDLC to cut full release verification from two or three days to under two hours, and how the SQA role changed along the way.","https://cdn.sqa.tw/articles/ai-sdlc-verification-bottleneck-cover.jpg","2026-11-05T12:00:00+08:00","perspective",{"id":27,"slug":27,"title":28,"excerpt":29,"image":30,"date":31,"category":25},"sme-guide-to-choosing-ai-agent","AI agents are everywhere. How should a small business choose one?","We were recently asked to run a full post-deployment review of an AI agent, and found it had quietly caused five times too much raw material to be ordered from customer order emails, and nearly created a liquidity risk on the finance side. The problem with agents is rarely that they are not smart enough. It is that nobody asked the right questions.","https://cdn.sqa.tw/articles/sme-agent-black-box-inspection.jpg","2026-09-01T15:00:00+00:00",{"id":33,"slug":33,"title":34,"excerpt":35,"image":36,"date":37,"category":25},"ai-coding-team-collaboration-three-patterns","Three patterns we see once AI coding reaches the team","After working with more than ten development teams over the past six months, we keep seeing the same thing: individuals getting faster with AI does not mean the team is ready to get faster together.","/images/blog/ai-coding-teams-editorial.jpg","2026-08-20T08:00:00+00:00",{"slug":39,"title":40,"date":41},"error-code-as-software-quality-indicator","[A hidden SQA indicator] Structured error codes","2025-12-21T05:16:00+00:00",{"slug":43,"title":44,"date":45},"pikmin-with-fake-gps","It turns out a testing tool can be for fun too","2026-02-03T12:44:00+00:00",[47,48],"en","zh",1794970112941]