v0.1.3

What's Changed

• fix: allow model downloads on mobile data by @artsaraiva in #426

Full Changelog: v0.1.2...v0.1.3

v0.1.2

What's Changed

• ci: pin actions to commit SHAs by @artsaraiva in #419
• docs: add Get it on Obtainium badge by @artsaraiva in #421
• feat: show page time in the Performance readout by @artsaraiva in #424

Full Changelog: v0.1.1...v0.1.2

v0.1.1

What's Changed

• chore: drop personal setup and stale docs by @artsaraiva in #413
• docs: add contributor guide and templates by @artsaraiva in #415
• feat: add the launcher icon by @artsaraiva in #417

Full Changelog: v0.1.0...v0.1.1

v0.1.0

What's Changed

• docs: forbid AI attribution in commits and PRs by @artsaraiva in #219
• refactor: collapse the catalog delivery tier (#189) by @artsaraiva in #220
• fix: keep the batch overflow fallback marker apart from the failure cause (#202) by @artsaraiva in #233
• test: pin id-keyed reply edges (#229) by @artsaraiva in #234
• docs: record that quality belongs to the model (#231) by @artsaraiva in #236
• fix: return generated text as bytes so invalid UTF-8 cannot abort the app (#235) by @artsaraiva in #237
• test: add report-only on-device speed benchmark (#230) by @artsaraiva in #238
• chore: delete the eval system (#232) by @artsaraiva in #239
• refactor: remove ML Kit and OPUS-MT floor engines by @artsaraiva in #240
• refactor: select detection and OCR models per slot by @artsaraiva in #241
• feat: split Settings into category sub-screens by @artsaraiva in #242
• feat: shared slot picker with pending selection by @artsaraiva in #243
• feat: setup slot screen and deleting the model in use by @artsaraiva in #244
• feat: reader-tunable bubble confidence threshold by @artsaraiva in #245
• feat: cancel an in-progress translation from the floating button by @artsaraiva in #246
• feat: resource readout and model resource limits by @artsaraiva in #247
• feat: quick-settings chip strip on Home by @artsaraiva in #249
• docs: add README with before/after demo page by @artsaraiva in #250
• feat: add Dependabot, CodeQL, and CodeRabbit config by @artsaraiva in #251
• ci: faster PR checks + llama.cpp 5266f24 bump by @artsaraiva in #260
• chore(deps): bump org.jetbrains.kotlin.jvm from 1.9.20 to 2.4.20 by @dependabot[bot] in #257
• chore(deps): bump actions/checkout from 6 to 7 by @dependabot[bot] in #259
• chore(deps): bump ml/llama.cpp from 263cc04 to 5266f24 by @dependabot[bot] in #252
• chore(deps): bump the gradle-minor-patch group across 1 directory with 20 updates by @dependabot[bot] in #253
• chore(deps): bump com.squareup.retrofit2:retrofit from 2.9.0 to 3.0.0 by @dependabot[bot] in #254
• chore(deps): bump com.squareup.okhttp3:okhttp from 4.12.0 to 5.5.0 by @dependabot[bot] in #255
• chore(deps): bump actions/setup-java from 5 to 6 by @dependabot[bot] in #256
• chore(deps): bump com.android.library from 8.5.2 to 9.4.0 by @dependabot[bot] in #258
• fix(build): revert incompatible Dependabot bumps by @artsaraiva in #261
• fix: typeset into the bubble interior by @artsaraiva in #268
• chore(deps): stop Dependabot proposing Kotlin 2-only bumps by @artsaraiva in #269
• fix(overlay): remove drag-to-close zone by @artsaraiva in #275
• fix: keep the Home quick-settings chips on one row by @artsaraiva in #276
• feat(overlay): compact the quick settings popup by @artsaraiva in #296
• docs: record ADR-0017 experimental curated tier by @artsaraiva in #297
• docs: record the Ternary-Bonsai spike (#299) by @artsaraiva in #303
• feat: render prompts with llama.cpp's Jinja engine by @artsaraiva in #304
• feat: add Qwen3.5 2B as experimental deliverable by @artsaraiva in #305
• feat(translation): per-model maker sampling by @artsaraiva in #307
• fix: reconcile READY models whose GGUF is gone by @artsaraiva in #309
• fix: honour reader threshold in ONNX parse by @artsaraiva in #310
• feat: speed label and Slow-deliverable warning in the model picker by @artsaraiva in #311
• feat: per-entry system message and Ministral 3 3B/8B by @artsaraiva in #312
• feat: add Gemma 4 E2B and E4B as Experimental deliverables by @artsaraiva in #313
• feat: add the rest of the Qwen Experimental line by @artsaraiva in #314
• feat: add Hy-MT2 1.8B as a per-line Experimental deliverable by @artsaraiva in #315
• feat: add Uncensored Experimental deliverables by @artsaraiva in #316
• feat: load runtime-selected ARM CPU backend variants by @artsaraiva in #318
• fix: repack weights only when a second copy fits by @artsaraiva in #321
• fix: unknown device RAM fits only the default by @artsaraiva in #322
• feat: add Ternary-Bonsai 4B and 8B as Experimental by @artsaraiva in #323
• feat: show art and a count for unanswered bubbles by @artsaraiva in #344
• feat: bubble card and swipe dismissal on the translated page by @artsaraiva in #342
• feat: stream bubbles on per-line deliverables by @artsaraiva in #346
• docs: record the reading-flow decisions by @artsaraiva in #347
• feat: stream batch reply lines as they decode by @artsaraiva in #348
• feat: show an already translated page instantly by @artsaraiva in #349
• feat: share the last page's report from quick settings by @artsaraiva in #354
• chore(deps): bump ml/llama.cpp from 5266f24 to 7fe450e by @dependabot[bot] in #356
• chore: turn off automatic CodeRabbit reviews by @artsaraiva in #359
• fix: keep taps outside quick settings from the app by @artsaraiva in #357
• fix: stop reading when screen sharing ends by @artsaraiva in #360
• feat: add a switch to turn page recall off by @artsaraiva in #361
• docs: research Jev-style bubble hints (#350) by @artsaraiva in #362
• fix: letterbox the page for the bubble detector by @artsaraiva in #371
• fix: keep the box that contains a kept bubble by @artsaraiva in #372
• fix: let the last line end after its translation by @artsaraiva in #373
• fix: decode OCR from the model's start token by @artsaraiva in #374
• fix: read panels from the page's gutters by @artsaraiva in #375
• fix: skip status bar and non-Japanese bubbles by @artsaraiva in #376
• docs: fix agent tool names, move standards, add page-report script by @artsaraiva in #383
• feat: language label in the model picker by @artsaraiva in #382
• test: cover TranslationPipeline bubble filtering by @artsaraiva in #385
• ci: enforce the mechanical code standards by @artsaraiva in #387
• fix: keep dropped OCR reads in the page report by @artsaraiva in #386
• fix: read and draw bubbles under a hidden status bar by @artsaraiva in #389
• feat: skip OCR reads the OCR model is unsure of by @artsaraiva in #390
• fix: show overlay failures on the status line by @artsaraiva in #392
• feat: reuse OCR decoder state between steps by @artsaraiva in #394
• feat: translate pages into the reader's target language by @artsaraiva in #395
• feat: keep target language and model choices compatible by @artsaraiva in #396
• docs: research register hints for target languages by @artsaraiva in #397
• docs: publish spike code as a tag, not a branch by @artsaraiva in #398
• feat: name glossary for consistent names by @artsaraiva in #400
• docs: license Yomu under AGPL-3.0 by @artsaraiva in #404
• build: publish a signed APK on GitHub Releases by @artsaraiva in #408
• docs: README launch info and bug-report form by @artsaraiva in htt...