{"data":{"slug":"purewhiter-mobilegym","name":"mobilegym","tagline":"Browser-hosted Android Simulator for Mobile GUI Agent Research","github_url":"https://github.com/Purewhiter/mobilegym","owner":"Purewhiter","repo":"mobilegym","owner_avatar_url":"https://avatars.githubusercontent.com/u/54441452?v=4","primary_language":"Python","stars":795,"forks":136,"topics":["agent","agents","ai","android","automation","benchmark","gym","llm","llm-agents","mobile-agent","online-rl","react","reinforcement-learning","rl","rl-environment","sim-to-real","simulator","typescript","vlm"],"archived":false,"github_pushed_at":"2026-08-28T14:09:15+00:00","maintenance_label":"Active","stars_delta_30d":40,"url":"https://www.graphcanon.com/tools/purewhiter-mobilegym","markdown_url":"https://www.graphcanon.com/tools/purewhiter-mobilegym.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/purewhiter-mobilegym","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=purewhiter-mobilegym","description":"[EMNLP 2026] MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research · 浏览器里运行的安卓模拟器 · Browser-hosted Android Simulator · Verifiable Evaluation · Scalable Online RL Training","homepage_url":"https://mobilegym.dev","license":"Apache-2.0","open_issues":9,"watchers":2,"ai_summary":"MobileGym is a platform focusing on verifiable evaluation and scalable online reinforcement learning training specifically for mobile GUI agents in an android simulation environment hosted in browsers.","readme_excerpt":"## 📦 Licensing\n\nMobileGym uses **two licenses** by design — please read both before redistributing.\n\n- 🛠️ **Code** → [`LICENSE`](LICENSE) — **Apache License 2.0**.\n  Core MobileGym source files (`os/`, `apps/`, `system/`, `bench_env/`, `scripts/`, `docs/`).\n- 🧪 **Training code** → [`mobilegym-rl/LICENSE`](mobilegym-rl/LICENSE) — **Apache License 2.0**.\n  Online RL pipeline under `mobilegym-rl/`; vendored third-party components retain their upstream notices / licenses where provided.\n- 📚 **Data & content** → [`LICENSE-DATA`](LICENSE-DATA) — **CC BY-NC 4.0**.\n  All replaceable JSON, synthetic / AI-generated content, simulated UGC and icons under `mobilegym-data/`, `apps/*/data/`, `apps/*/assets/`. **Non-commercial academic use only.**\n\nThe split exists because we want the *platform code* to be permissively reusable while the *content* (which includes derived representations of third-party brands for research realism) remains scoped to research. See [DISCLAIMER.md](DISCLAIMER.md) for the full story.\n\n<br/>","github_created_at":"2026-05-14T11:09:52+00:00","created_at":"2026-07-15T10:56:45.686124+00:00","updated_at":"2026-09-20T04:55:41.207345+00:00","categories":[{"slug":"ai-agents","name":"AI Agents","url":"https://www.graphcanon.com/categories/ai-agents","markdown_url":"https://www.graphcanon.com/categories/ai-agents.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/ai-agents"},{"slug":"model-training","name":"Model Training","url":"https://www.graphcanon.com/categories/model-training","markdown_url":"https://www.graphcanon.com/categories/model-training.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/model-training"}],"tags":[{"slug":"agent","name":"agent"},{"slug":"ai","name":"ai"},{"slug":"android","name":"android"},{"slug":"automation","name":"automation"},{"slug":"benchmark","name":"benchmark"},{"slug":"gym","name":"gym"},{"slug":"llm-agents","name":"llm-agents"},{"slug":"mobile-agent","name":"mobile-agent"}],"trust":{"provenance":{"is_fork":false,"github_id":1238694393,"owner_type":"User","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-20T04:55:38.733Z","maintenance":{"label":"Active","score":82,"methodology":"github_public_v1","releases_90d":2,"days_since_push":22,"last_release_at":"2026-06-26T18:31:05Z","stars_delta_30d":40,"open_issues_delta_30d":2},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-15T10:56:46.860Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"mcp":{"source":"repo_scan","observed_at":"2026-09-20T04:55:39.754Z","server_manifest":false},"scan":{"source":"repo_scan","observed_at":"2026-09-20T04:55:39.754Z"},"languages":{"value":["python","javascript","typescript"],"source":"github.language+package.json","observed_at":"2026-09-20T04:55:39.754Z"},"license_spdx":{"value":"Apache-2.0","source":"github.license","observed_at":"2026-09-20T04:55:39.754Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["You require a browser-based solution to simulate Android environments for testing and training mobile graphical user interface (GUI) agents.","Need a scalable platform for online reinforcement learning that supports verifiable evaluation of your AI models specifically tailored for mobile applications."],"when_not_to_use":["Your project does not involve mobile GUI agent research or development; MobileGym's specialty is in this specific application and may not be useful for general-purpose simulation。","You are looking for a solution that runs exclusively on local servers rather than browser-hosted environments."],"source":"enrich:decision_facts","observed_at":"2026-07-17T11:51:57.100Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"MobileGym specializes in browser-hosted Android simulation for reinforcing learning training of mobile GUI agents. It is licensed under Apache-2.0 and written in Python."}]}}