This page brings together the projects I maintain, am developing, have made public, or have not yet released. They include Android performance analysis, Perfetto tools, AI automation, iOS apps, Android demos, test suites, blog series, communities, and my accounts on other platforms.
If this is your first visit, start with what you need: the Perfetto and Systrace series for Android performance analysis; SmartPerfetto, TraceFix, and Android App Memory Analysis for tools; OpenClaw and AI Field Notes for AI-assisted knowledge management and everyday work; or the links at the end to contact me.
The projects fall into four areas. Performance analysis includes tools such as SmartPerfetto, Android App Memory Analysis, TraceFix, and Perfetto Auto-Pin, alongside the High Performance Friends Circle test platform. AI and automation includes OpenClaw, AI Field Notes, Gracker Skills, and Open Design. Personal apps include the iOS projects 100Years and iBattery, and the Android project Juju. The main content projects are this blog and Android Weekly. The entries below describe their status, including ongoing maintenance, current development, and paused work.
The project overview is also part of the blog index and personal projects. This original page remains available for reference.
Update note: The Chinese source was reviewed on October 14, 2026; this English translation was completed on October 15. The original project scope and personal directions are preserved, with technical limits, public links, and version changes added during review. Personal usage accounts retain their original time context. Describing a private app under development does not establish that it has shipped or is available in an app store.
Main Directions
- Android system performance and stability: my daily work and long-term writing focus on Android Framework, APM, Perfetto, Systrace, startup, jank, ANRs, memory, and power consumption.
- Perfetto and trace-analysis tools: turning experience with manual trace analysis into tools, scripts, Skills, and reusable workflows.
- AI agents and personal automation: using OpenClaw, Claude Code, Codex, Obsidian, Telegram, and scheduled tasks for research material, daily reports, knowledge bases, writing, and engineering collaboration.
- Personal apps: building small iOS and Android products around real needs.
- Technical writing and research collections: maintaining Android Performance, Android Weekly, AI Field Notes, and several public Skills.
Public Projects
SmartPerfetto
SmartPerfetto is the open-source project receiving most of my attention. It adds an AI Assistant to Perfetto: after loading a trace, users can ask questions in natural language. The backend gathers evidence through trace_processor_shell, SQL, YAML Skills, and scenario strategies before producing an analysis.
Its main parts are:
- A Perfetto UI fork, Gracker/perfetto, which hosts the AI Assistant plugin.
- A TypeScript backend for agent execution, MCP tool calls, Skill execution, report generation, and streaming output.
- YAML Skills and scenario strategies that turn analysis experience with scrolling, startup, ANRs, Binder, CPU, memory, and rendering into executable units.
- Related articles: open-source introduction, architecture deep dive, and architecture Q&A.
SQL and Skills provide traceable data; the model interprets it. A report or citation does not mean every claim has passed verification. The architecture articles retain their own version context; use the documentation for the relevant release when installing or configuring the tool.
Android App Memory Analysis
Android-App-Memory-Analysis brings together data such as dumpsys meminfo, dumpsys gfxinfo, HPROF, and smaps for Android app memory analysis.
It is intended for these tasks:
- Collecting memory-related data for a target app when ADB is authorized and permissions allow it. HPROF and smaps availability depends on app debugging properties, system permissions, and device policies.
- Correlating Java heap, native memory, graphics memory, and process-level data. The sources use different accounting conventions, so their numbers cannot simply be added together; missing values are not zero.
- Producing readable reports to reduce manual switching between commands and files.
- The public documentation now covers Android 17 / API 37 adaptation. Demo 1.1.0 uses compile/target SDK 37 and retains the Android 16, 16 KB page-size, and edge-to-edge baseline. Compatibility with a particular device still requires actual collection and validation.
TraceFix
TraceFix is an Android bytecode-instrumentation plugin that inserts android.os.Trace.beginSection/endSection around methods at build time.
It reduces the time and omissions involved in adding trace sections manually. The sections can then be inspected in Perfetto. The published Maven Central artifact is io.github.gracker:TraceFix:0.1.0.
As of October 14, 2026, Maven Central metadata still lists 0.1.0. The source README describes 0.2.x for Android 17 / AGP 9.1.x, which does not establish that 0.2.0 has been published to Central. Instrumentation also needs a suitable capture configuration and trace tool before sections become visible. Check the compatibility matrices for older artifacts and newer source separately.
Perfetto Auto-Pin
Perfetto-AutoPin-For-Chrome is a Chrome extension that automatically pins frequently used tracks in Perfetto UI.
It includes presets for Frame Analysis, Startup Analysis, Input Response, Memory Analysis, Binder Calls, SurfaceFlinger, SystemUI, CPU, and other scenarios, as well as custom track collections. The aim is to reduce repeated expanding and pinning after opening a trace.
The extension depends on Perfetto UI’s internal window.app interface, so compatibility can change with the UI version. A preset name does not guarantee that a trace contains the corresponding tracks. Check the extension and UI versions before use.
High Performance Friends Circle
Friends-Circle-Demo-Apks-For-Power-and-Performance-Test is an Android performance and power test platform built around a social-feed interface.
The repository is organized as a workload test suite:
- Common scenarios include startup, scrolling, and page transitions.
- It includes multiple UI technology stacks and workload intensities.
- It supports investigating the main thread, RenderThread, SurfaceFlinger, CPU, power consumption, and frame behavior.
- There is also a Flutter version.
Systrace / Perfetto Utilities and Historical Demos
These are long-running tools and demos accompanying older articles. They remain learning and reproduction material; they are not guaranteed to run unchanged with current Android or Gradle versions. Check system-signature requirements, private interfaces, and capture formats against each project’s original version.
- SystraceForBlog: Systrace attachments and case material used in blog articles.
- SystraceAnalysis: an early systrace / ftrace analysis tool.
- android_cpu_usage_calculation: scripts related to Android CPU usage calculation.
- FPSDump: an FPS demo depending on old private SurfaceFlinger interfaces and a system signature, not a general frame-rate capture method for modern apps.
- GameBench: an Android game-performance tool with server and client components.
- Android-FPS-Calculation: an Android FPS calculation demo.
- Android_HardwareLayer_Example: a hardware-layer animation performance example.
- DelayLoadSample: an Android DelayLoad example.
- ListHeaderAnimationPerformance: a list-header animation performance test.
- AndroidLunchModeTest: an Android launch-mode test demo.
- AndroidUML: Android UML material organized with PlantUML.
Historical Android App / Java Projects
These are mainly learning projects, article examples, and early tools:
- Cats: Kotlin practice and a personal app demo.
- Girls: a Gank client practice project.
- Flutter-Gank: a Flutter practice project.
- FriendsCircle_Flutter_Demo: a Flutter social-feed demo.
- SlideDelete-ListView: a ListView swipe-to-delete example.
- Android-NumberLimitEditText: an EditText with input-length limits.
- SystemMonitor: an Android system-monitoring example.
- JavaReflect: Java reflection tests.
- AndroidInterview-PrintAllViews: a View-tree printing demo.
- MemoryAnalysisTools: memory-analysis tools for early Meizu devices.
- Fragment-NetEase: an early Fragment framework, now archived.
Collections, Sites, and Translations
Beyond tools and demos, I have maintained several collections, sites, and translation projects:
- Android Internal Wiki: a public Android system knowledge base, with a website and introduction.
- Android Performance Ecosystem: a navigation point for performance tools, knowledge, and example projects.
- awesome-android-ai-dev-sources: technical people, blogs, and development resources I follow or collect.
- Gracker.github.io: the Android Performance deployment repository.
- hexo-theme-reading: an early Hexo blog theme.
- claude-code-from-source-zh-cn: the Simplified Chinese translation of Claude Code from Source.
See my GitHub profile for the complete public repository list. This page focuses on projects most closely related to my long-term work, blog, tools, and apps.
SmartProtobuf
SmartProtobuf is a tool project concerned with Protobuf / trace data structures. Once the public information is ready, I will add its repository, usage instructions, and related articles here. Until then, I am leaving out unstable addresses.
AI and Automation Projects
OpenClaw
OpenClaw is a local AI automation system I have been using extensively. It connects Telegram, Obsidian, scheduled tasks, local and cloud models, and engineering tools for information organization, knowledge-base maintenance, daily reviews, GitHub monitoring, close reading of papers, writing-material archives, and engineering collaboration.
Related articles:
- After Three Weeks Running OpenClaw Locally, I Realized It Wasn’t a Chatbot
- OpenClaw FAQ: Tokens, Capabilities, Local Models, Privacy, and Practical Use
AI Field Notes
awesome-ai-field-notes is an AI research collection. Its documentation describes an OpenClaw-based collection and site-generation workflow. The original site, godofgpt.com, could not be reached normally during the October 14, 2026 review, so GitHub is the reference entry point here.
The workflow collects, deduplicates, categorizes, scores, translates, and generates pages. It covers models, agents, AI coding, infrastructure, industry observations, and learning resources. For me, it is a long-term archive of the AI information stream and a public example of OpenClaw automation.
Gracker Skills
I have packaged some of my writing, research, and diagram workflows as public Skills:
- gracker-writing: technical articles, practical tool retrospectives, FAQs, methodology, and long-form WeChat articles.
- gracker-deep-research-skill: a three-stage research workflow separating local discovery, evidence collection, and offline writing.
- gracker-diagrams-skill: technical diagrams, architecture diagrams, and infographics. It prepares a prompt pack by default; actual image generation depends on the environment’s drawing capabilities.
Open Design
Open Design is a design-tool project I previously forked and helped maintain. It explored bringing local coding agents, design systems, Skills, export, and preview into one workflow. The original public repository returned 404 during the October 14, 2026 review, so no installation link is provided here.
Apps in Development
100Years
100Years is an iPhone / Apple Watch app for health trends and long-term health signals. It combines HealthKit, lifestyle, medical-checkup records, activity, and sleep data. Deterministic models produce estimates, while AI explains them. Lifespan, centenarian, and healthspan values are product-model estimates; they should not be treated as externally clinically calibrated personal survival probabilities or medical conclusions.
100Years is still being refined, and its source repository is private. Development covers data import, medical indicators, AI explanations, HealthKit synchronization, a SwiftUI design system, localization, and prerelease checks.
Public support, privacy, and terms material is available in the 100 Years Life support repository. This is a support-material link only.
iBattery
iBattery monitors equipment battery levels for cyclists. The iOS version covers Bluetooth sports equipment such as power meters, heart-rate straps, electronic shifting systems, cadence sensors, trainers, and cycling computers.
Its main functions include:
- Automatically discovering and classifying nearby Bluetooth sports equipment.
- Updating battery status and issuing reminders when a device exposes readable battery data. Detecting or identifying a device does not guarantee access to its battery level.
- Showing the most recent battery reading for frequently used equipment in iOS widgets. Widget refresh and Bluetooth sampling are separate, so the reading’s timestamp matters.
- Supporting multiple languages, local caching, notifications, and themes.
The iOS repository is private. An Android version is also being developed locally to bring the iOS experience to Android.
Juju
Juju is an Android app for my family, recording a child’s age, milestones, photos, and home-screen widgets. It is intended for private use and is not being promoted as a public open-source project.
Content Projects
Android Performance Blog
Android Performance is the public content site I have maintained the longest.
The main topics are:
- Blog index: the main article directory.
- Perfetto: tools, capture, UI, SQL, Trace Processor, Binder, CPU, Vsync, RenderThread, and automated analysis.
- Systrace: Android mechanisms, thread states, Vsync, Input, SurfaceFlinger, Binder, Triple Buffer, and SystemServer.
- Smoothness and responsiveness: startup, scrolling, dropped frames, low memory, CPU, and Runnable / Running / Sleep states.
- ANRs, memory, Framework, app development, reading notes, recommendations, and Zhihu Q&A.
Android Weekly
Android Weekly is my free technical newsletter covering Android, Linux, iOS, Flutter, performance optimization, and engineering tools.
Newsletter links:
The Performance Community (Closed)
The Performance was my Android performance community. It stopped operating in 2026 and no longer accepts new members. Its material on analysis fundamentals, tools, and case studies will gradually be organized into the blog and Android Internal Wiki.
Platform Links
- Blog: Android Performance
- GitHub: Gracker
- Zhihu: Gracker
- Zhihu column: Android Performance
- Juejin: Gracker
- Juejin column: Android Weekly
- X / Twitter: Gracker_Gao
- Weibo: gracker520
- RSS: Subscribe
- Email: Dreamtale.jg@gmail.com
For WeChat and discussion groups, see About Me, where contact details are maintained.
How to Use This Index
- To study Android performance systematically, start with the blog index.
- To analyze Perfetto traces with AI, see SmartPerfetto and its open-source introduction.
- For app memory analysis, see Android-App-Memory-Analysis.
- To instrument apps automatically, see TraceFix.
- For performance-test demos, see High Performance Friends Circle.
- For AI automation, see the OpenClaw practice article and the AI Field Notes collection.
- To contact me, use About Me or the platform accounts above.