v0.14.0

What's Changed

New Features 🎉

• feat: migrate android code to @midscene/android by @quanru in #505
• feat: use aiTap(prompt, opt) style interface by @yuyutaotao in #512
• feat: print stats in debug by @yuyutaotao in #515
• feat: update deepThink interface by @yuyutaotao in #517
• feat(llm): allow setting context for aiAction by @yuyutaotao in #528

Bug Fixes 🐞

• fix(report): resolve report script error by @zhoushaw in #516
• fix: force screenshot when the app doesn't allow it by @quanru in #518
• fix: ci testing by @yuyutaotao in #529

Document 📖

• docs: updat faq by @yuyutaotao in #501
• docs: lazy loading report.gif by @chenjiahan in #503

Other Changes

• chore: not run android tests in ci && update tests for android settig case by @quanru in #502
• chore: support new ui-tars model json format by @zhoushaw in #506
• chore(report): init report project by @zhoushaw in #511
• refactor(report): make the report page an app type to improve the development experience by @zhoushaw in #513
• refactor(logger): use log file output instead of command line output in Node by @zhoushaw in #509
• chore: delete unless visualizer logic by @zhoushaw in #514
• refactor(report): optimize visualizer common logic by @zhoushaw in #519
• refactor(visualizer): optimize extension logic by @zhoushaw in #525
• refactor: android api by @quanru in #521

Full Changelog: v0.13.1...v0.14.0

v0.13.0

What's Changed

New Features 🎉

• feat: export overrideAIConfig in chrome extension by @zhoushaw in #482
• feat: enable search area for locate by @yuyutaotao in #473
• feat(playwright): add aiTap/aiInput and other AI actions for Playwright integration by @zhoushaw in #489
• feat: update prompt for qwen by @yuyutaotao in #491

Full Changelog: v0.12.8...v0.13.0

v0.12.8

What's Changed

New Features 🎉

• feat: faster screenshotBase64 by @VinceOPS in #449
• feat: optimize locator by @yuyutaotao in #456

Bug Fixes 🐞

• fix(web-integration): resolve waitForNavigation timeout by @zhoushaw in #468

Other Changes

• chore: vince ops feat/faster screenshot base64 in web integration by @yuyutaotao in #450
• refactor(chrome-devtool): extract the chrome-devtool logic into an application, support local development debugging, and add contribution guidelines. by @zhoushaw in #476
• chore: fix lint error by @zhoushaw in #477

New Contributors

• @VinceOPS made their first contribution in #449

Full Changelog: v0.12.7...v0.12.8

v0.12.7

What's Changed

New Features 🎉

• feat(core): fix MIDSCENE_CACHE=true not work by @Peterwub in #452

Bug Fixes 🐞

• fix: enhance CoT of qwen, make it more stable across running by @yuyutaotao in #446

New Contributors

• @Peterwub made their first contribution in #452

Full Changelog: v0.12.4...v0.12.7

v0.12.4

What's Changed

New Features 🎉

• feat(core/ui-tars): support aiAssert() & aiWaitFor for UI-Tars model by @xuanzhaopeng in #412
• feat(chrome-devtool): allow longer connection of chrome bridge by @yuyutaotao in #416
• feat(yaml): Add cache support for yaml files, allowing file-level caching to improve execution speed and stability. Cache ID is automatically set to yaml filename. by @zhoushaw in #430

Bug Fixes 🐞

• fix: print warning for oversized input by @yuyutaotao in #414
• fix(core): the prompt of qwen by @yuyutaotao in #427
• fix: over planning for Qwen in page with form by @yuyutaotao in #429

Document 📖

• docs: optimize agent api doc by @zhoushaw in #415
• docs: optimize bridge api structure by @zhoushaw in #417
• docs: update faq by @yuyutaotao in #422
• docs: fix markdown content of quick-experience by @yuyutaotao in #424
• docs: update instruction for bridge mode by @yuyutaotao in #428

Other Changes

• chore: potential fix for code scanning alert no. 13: Clear-text logging of sensitive information by @zhoushaw in #423
• chore: optimize lint error by @zhoushaw in #425
• chore: rename core package by @yuyutaotao in #431
• chore: tuning the prompt of qwen by @yuyutaotao in #432
• chore: tuning the prompt of qwen by @yuyutaotao in #433

New Contributors

• @xuanzhaopeng made their first contribution in #412

Full Changelog: v0.12.0...v0.12.4

v0.12.0

What's Changed

New Features 🎉

• feat: qwen2.5 vl model by @yuyutaotao @zhoushaw in #383

Bug Fixes 🐞

• fix: partial scroll issue in chrome extension by @yuyutaotao in #401
• fix: coord bias of Qwen model by @yuyutaotao in #407
• fix: ci of qwen model by @yuyutaotao in #410

Other Changes

• Potential fix for code scanning alert no. 14: Clear-text logging of sensitive information by @yuyutaotao in #405

Full Changelog: v0.11.3...v0.12.0

v0.11.3

What's Changed

Bug Fixes 🐞

• fix(report): resolve the report generation failure caused by the absence of a selector by @zhoushaw in #392
• fix(report): reduce playwright report size by @zhoushaw in #396

Other Changes

• doc: badge links by @cwtuan in #393

New Contributors

• @cwtuan made their first contribution in #393

Full Changelog: v0.11.2...v0.11.3

v0.11.2

What's Changed

New Features 🎉

• feat(yaml): export yaml runner in javascipt by @yuyutaotao in #368
• feat(browser): Add the forceSameTabNavigation configuration to prevent AI from opening new pages during operations, thus avoiding task interruptions. by @zhoushaw in #389
• feat(reporting): Add test ID to test data attributes by @myselfsaurabh in #380

Bug Fixes 🐞

• fix: upgrade langsmith sdk by @yuyutaotao in #374
• fix: correctly collect elements in absolute container by @yuyutaotao in #373
• fix: update discord links #375 by @yuyutaotao in #379
• fix(keybord): resolve selectAll/Copy/Paste event by @zhoushaw in #384
• fix(browser-event): optimize limit open new tab script and add yaml test by @zhoushaw in #391

Full Changelog: v0.11.0...v0.11.2

v0.11.0

What's Changed

New Features 🎉

• feat(web-extract): extract web content as a tree by @yuyutaotao in #337
• feat(cache): supports ui-tars model caching capability by @zhoushaw in #361

playwright

- playwright test --config=playwright.config.ts
+ MIDSCENE_CACHE=true playwright test --config=playwright.config.ts

puppeteer

- tsx demo.ts 
+ MIDSCENE_CACHE=true tsx demo.ts

const mid = new PuppeteerAgent(originPage, {
  cacheId: 'puppeteer-swag-sab)', // Add cache id
});

Bridge Mode

- tsx demo.ts
+ MIDSCENE_CACHE=true tsx demo.ts

const agent = new AgentOverChromeBridge({
  cacheId: 'star-midscene-github', // Add cache id
});

Effect

After enabling the cache, the execution time is significantly reduced, for example, from 1m16s to 23s.

before

after

• feat: use different color for annotations by @yuyutaotao in #366

Bug Fixes 🐞

• fix: improve waitForNetworkIdle with timeout handling by @myselfsaurabh in #365
• fix(ui-tars): resolve page down and page up event error by @zhoushaw in #370
• fix: ai evaluation by @yuyutaotao in #371

Other Changes

• doc: enable code search in rspress by @yuyutaotao in #356
• workflow: enhance error messages and improve the debugging experience. by @zhoushaw in #363
• chore: optimize e2e test by @zhoushaw in #364

New Contributors

• @myselfsaurabh made their first contribution in #365

Full Changelog: v0.10.5...v0.11.0

v0.10.5

What's Changed

Bug Fixes 🐞

• fix(prompt): resolve params format error by @zhoushaw in #343

Full Changelog: v0.10.4...v0.10.5