Midscene
Vision-driven UI automation for web, mobile, and desktop — no selectors needed.
If you're tired of brittle selectors and need cross-platform UI automation, Midscene is a smart, free, open-source pick. Its multi-model approach and strong benchmarks (93.1% AndroidWorld) back its claims. It's ideal for self-hosters and scripters, but skip it if you want a managed SaaS or deterministic element selection. For managed alternatives, consider TestIM, Mabl, or Applitools.
Verified 6d ago · liveness 77/100 · cite: rightaichoice.com/tools/midscene
- QA engineers writing cross-platform UI tests that break less often
- Developers automating canvas, native apps, or cross-origin frames
- Teams wanting selector-free test stability without cloud costs
- AI agent developers needing vision-based web/app testing
- Teams needing a fully managed SaaS test platform
- Users requiring pure text-based automation without vision
- Complete beginners unfamiliar with test runners
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Midscene if you need a fully managed test platform with cloud orchestration, real device farms, or zero infrastructure — or if you have no experience with test runners and AI APIs, since you'll be self-hosting and wiring up your own model keys.
You must provide your own API keys for cloud models like Doubao-Seed-2.1-turbo, qwen3.7-plus, or gemini-3.5-flash — costs scale with your test volume and can add up with large suites.
Midscene is completely free, MIT-licensed, and open-source — there are no per-test, per-seat, or per-device fees. Compared to commercial tools like TestIM or Mabl (which charge per test run or device), Midscene is massively cheaper if you have the engineering capacity to self-host. For small teams with modest test volumes, the cost of model API calls is often negligible.
In short
Midscene — Vision-driven UI automation for web, mobile, and desktop — no selectors needed. Best for QA engineers writing cross-platform UI tests that break less often, Developers automating canvas, native apps, or cross-origin frames, Teams wanting selector-free test stability without cloud costs. Free to use.
What's new in Midscene
Checked 6 days agoAcross the latest 5 updates: 5 feature updates.
Midscene v1.10 released with BDD-style Gherkin scripts, Doubao-Seed-2.1 support, and MCP retirement
v1.10 adds runGherkinScenario for JS/TS and YAML flows, upgrades Doubao-Seed-2.1-turbo as recommended model, and retires all MCP server packages.
Kimi and Xiaomi MiMo model support added
v1.9 expands model support with Kimi and Xiaomi MiMo, broadening options for localization.
AndroidWorld benchmark report published
Midscene achieves Pass@1 93.10%, Pass@2 95.69%, Pass@3 97.41% on AndroidWorld with v1.9.5.
Doubao-Seed-2.1-turbo now recommended model
Doubao-Seed-2.1-turbo becomes recommended for fast localization. Family set via MIDSCENE_MODEL_FAMILY=doubao-seed.
MCP server packages retired in v1.10
All @midscene MCP packages removed. Users relying on MCP should pin Midscene to 1.9.8.
What people actually say about Midscene — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
19 mentions across 3 sources (Hacker News, YouTube, GitHub) · researched Aug 2, 2026.
- +Eliminates reliance on brittle CSS selectors and XPaths with vision-based detection.
- +Supports web, mobile, and desktop, covering a wide range of platforms.
- +Integrates with popular tools like Playwright and Puppeteer for easy adoption.
- +Offers a Chrome extension for quick experimentation and demos.
- +Free and MIT-licensed, making it accessible for individual developers and enterprises.
- −Configuration for advanced setups (Azure OpenAI) is complex and error-prone.
- −Limited community support; few active discussion forums or quick help channels.
- −Requires an understanding of AI models to optimize performance.
- −May not be suitable for non-programmers due to technical setup and scripting.
- −Vision-based approach may not work offline without API access to AI models.
- • External AI model API costs if using cloud services like OpenAI or Gemini, which are not included in the free tool.
Viability Score
How well maintained and how widely used is Midscene? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: August 2026
How we score →Key Features
- Vision-driven element detection from screenshots
- aiAct, aiLocate, aiAssert APIs for scripting
- Unified API across web, PC, Android, iOS, HarmonyOS
- Natural language test automation
- Multi-model strategy: plan + vision
- YAML script runner for workflow automation
- Gherkin BDD support (v1.10, Beta)
- Bridge Mode to drive desktop Chrome
- Chrome extension for quick experiments
- Visual replay report and interactive playground
- AI coding agent integration via Skills and CLIs
- Support for self-hosted open-source models
- Caching of AI planning and location results
- MCP packages retired in v1.10
- Supports file upload in Bridge Mode
About Midscene
Midscene is an open-source, MIT-licensed UI testing and automation toolkit that uses pure vision to control web, mobile, and desktop applications from natural language commands. You write scripts in JavaScript/TypeScript, YAML, or Gherkin (beta), and Midscene's AI handles the rest — locating elements, planning actions, and asserting outcomes — all from screenshots, so you never maintain fragile CSS selectors or XPath locators. It targets QA engineers, developers, and AI agent builders who need cross-platform test stability, especially for hard-to-reach UI like canvas drawings, native app views, and cross-origin frames. Midscene provides a unified API and test suite across web (Playwright, Puppeteer, Bridge Mode), mobile (Android, iOS, HarmonyOS), and desktop (macOS, Windows, Linux), with core methods like aiAct, aiLocate, and aiAssert. It pairs a planning model with a vision model for higher task completion rates, and supports a growing list of models: Doubao-Seed-2.1-turbo (recommended), qwen3.7-plus, gemini-3.5-flash, Kimi, and Xiaomi MiMo, plus self-hosted open-source options. The project includes a visual replay report and an interactive playground for debugging, and AI coding agents can drive tests via Midscene Skills and CLIs. Backed by ByteDance with 14k+ GitHub stars, Midscene reports strong benchmarks: 93.1% Pass@1 on AndroidWorld and 78.6% Pass@1 on MobileWorld. Note: Midscene is not a hosted SaaS — you self-host the SDK and run your own infrastructure. If you want a cloud orchestration layer or real device cloud, this isn't it. But for teams comfortable running their own tooling, Midscene is a compelling, free, selector-free choice.
Behind the Verdict
Midscene is a game-changer for teams that have struggled with selector-based testing, especially on modern web apps with canvas, shadow DOM, or cross-origin frames. The vision-first approach means you write natural-language scripts that read like plain English, and the AI handles element location. This dramatically lowers maintenance when UI changes — you don't rewrite locators, you just re-run the test. We love that it's fully open-source (MIT) and free, with no per-test or per-seat costs, which is rare among UI testing tools. The unified API across web, mobile, and desktop is a major differentiator. You write the same code to drive a browser, an Android emulator, an iOS simulator, and a desktop app. That's unusual and valuable for teams with cross-platform products. The Gherkin BDD support (beta) makes tests readable by non-developers, and the YAML script runner is great for CI pipelines. The multi-model strategy is smart — you can pair a planning model with a vision model to improve success rates, and you can swap models as pricing or quality changes. We tested with Doubao-Seed-2.1-turbo and it felt fast and reliable. But there are real tradeoffs: 1. **Performance depends on your AI model** — you need API keys (or self-hosted models), and every action incurs AI latency. In large suites, this can be slower than traditional selectors. 2. **It's not managed** — you must run the SDK, manage infrastructure, and handle model costs. No built-in CI cloud, no device farm. 3. **MCP integration was retired in v1.10** — if you relied on MCP servers, pin to 1.9.8, or move to Skills and CLIs. 4. **Gherkin support is beta and limited to single-Scenario** — not yet ready for complex BDD suites. Where does it fit? Midscene is ideal for: - Teams already using Playwright/Puppeteer that want to add vision-based steps for flaky areas. - Mobile app teams that want to test Android, iOS, and HarmonyOS with the same script. - AI agent developers who want their agents to test their own generated UI. - Open-source enthusiasts who want a self-hostable, free alternative to commercial tools. Where it doesn't fit: - Teams that need a managed test cloud with parallel execution, real devices, and zero infra. - Teams with no experience with test runners or AI APIs. - Teams that require 100% deterministic control (e.g., exact pixel coordinates) — Midscene is probabilistic by nature. Bottom line: a powerful, cutting-edge tool that puts the power in your hands, but demands you take responsibility for the plumbing. If you're ready for that, it's one of the best free options out there.
Researching Midscene? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Midscene actually fits — and what changes day-one when you adopt it.
You need to add regression tests for a canvas-heavy web app that keeps breaking when UI changes.
Outcome: You write a YAML script with natural language steps, run it against your existing Playwright setup, and get a visual replay report. Tests stay stable despite UI changes, and you save hours of selector maintenance.
You want to test your Android app on an emulator and your team also needs iOS coverage.
Outcome: You use Midscene's unified API to write one test that runs on both Android and iOS emulators, using aiAct to tap, swipe, and type in natural language. You get a single test suite that works across platforms, cutting duplication.
Your coding agent generates UI code and you need it to verify the UI works before merging.
Outcome: You add Midscene Skills to your agent's workflow, so it can run aiAssert checks on the new UI. The agent gets automated feedback, reducing manual verification time and catching regressions early.
Use Cases
- Automate end-to-end tests for a cross-platform mobile app using natural language scripts.
- Perform visual regression testing on a Canvas-heavy web application without maintaining CSS selectors.
- Create a BDD test suite that product managers can read and review using Gherkin scenarios.
- Integrate Midscene with an AI coding agent to let it test its own generated UI code.
- Run desktop app automation on Windows/Mac/Linux with a single unified API.
- Build a reusable YAML automation pipeline for different test environments.
- Test HarmonyOS apps with the same YAML workflow as web and Android.
- Use Bridge Mode to automate a browser that's already in use by a human.
Models Under the Hood
as of 2026-08-19
Limitations
- Midscene is an open-source library that requires local integration with your own API keys or self-hosted models.
- Performance depends on the chosen model, and BDD support is in Beta, limited to a single Scenario subset.
- MCP server packages have been retired.
- Because it relies on AI, results can be non-deterministic.
as of 2026-08-18
Verification history
We have re-verified Midscene 6 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Midscene tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Open Source
$0
Ideal for
Developers and QA teams who want a free, self-hosted UI automation toolkit and are comfortable with Node.js and managing their own model API keys.
What this tier adds
This is the only tier — MIT-licensed, $0, with the full feature set including YAML/Gherkin scripts, multi-platform support, and visual reports. No paid upgrades exist.
Where the pricing makes sense
The company stage and team size where Midscene's pricing actually pencils out — and where peers do it cheaper.
Midscene is completely free, MIT-licensed, and open-source — there are no per-test, per-seat, or per-device fees. Compared to commercial tools like TestIM or Mabl (which charge per test run or device), Midscene is massively cheaper if you have the engineering capacity to self-host. For small teams with modest test volumes, the cost of model API calls is often negligible.
Setup time & first value
How long it actually takes to get something useful out of Midscene — broken out by persona, not the marketing-page minute.
For a developer familiar with Node.js and Playwright: install the SDK, connect your AI model API key, and write a first test in under an hour. For a QA engineer without coding background, expect a few hours to get comfortable with YAML scripts and running reports. Desktop/mobile setup takes longer – plan half a day to configure device bridges and emulators.
Switching to or from Midscene
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From Playwright/Puppeteer selectors: Add Midscene's aiAct/aiAssert to your existing tests for flaky steps, gradually replacing selectors.
- ↗To a managed platform: Export your test steps as natural language and manually recreate them in TestIM or Mabl – there's no direct conversion tool, but your YAML/Gherkin scripts are portable as documentation.
Integrations
Resources & Guides
Tutorials & Learning
Official links
Tools that pair well with Midscene
Common stack mates teams adopt alongside Midscene, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Midscene vs Locus Robotics
Locus Robotics and Midscene serve completely different domains — physical warehouse automation vs. software UI testing. Choose Locus if you need proven AMRs for high-volume fulfillment with minimal facility redesign. Choose Midscene if you want free, selector-free, vision-driven automation for web/mobile/desktop apps. They aren't direct competitors; your use case dictates the choice.
Midscene vs Presto Voice
Choose Presto Voice if you run a QSR chain with drive-thrus and need a proven voice AI to boost revenue via upselling (Dairy Qeuen adoption validates enterprise readiness). Choose Midscene if you're a developer or QA engineer seeking free, open-source, vision-based UI automation that works across platforms—especially valuable for testing unlabeled or native elements. They solve completely different problems.
Midscene vs Truleo
For law enforcement agencies drowning in siloed data, Truleo is a purpose-built AI command center that slashes case research and report writing time. For software teams tired of brittle UI selectors, Midscene offers a free, open-source vision-driven test automation platform that works across web, mobile, and desktop—no subscriptions, no lock-in. Choose based on your domain: public safety or software quality.
Alternatives to Midscene
View allFrequently Asked Questions
Used Midscene? Help shape our editorial sentiment research.