{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:G22DYMOXRRSU6XP7HJC4RSWBPN","short_pith_number":"pith:G22DYMOX","schema_version":"1.0","canonical_sha256":"36b43c31d78c654f5dff3a45c8cac17b6b462e7cd1d3382cc86219411e699239","source":{"kind":"arxiv","id":"2607.28645","version":1},"attestation_state":"computed","paper":{"title":"Looks Right, Works Right: A Project-Level Benchmark for Multi-Screen Mobile App Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.HC","authors_text":"Cuiyun Gao, Fan Wu, Qing Liao, Yang Xiao, Yiming Huang, Yujia Chen","submitted_at":"2026-05-29T05:48:19Z","abstract_excerpt":"Recent multimodal large language models can convert visual designs directly into executable code, but real mobile products require multiple screenshots to become a buildable codebase with shared components and working navigation. This project-level setting exposes three limits of existing design-to-code benchmarks: they focus on single-page generation rather than complete codebases, cannot evaluate cross-page navigation, and do not measure project-wide maintainability. We introduce MobileForge, the first benchmark for project-level multi-screen mobile app generation, comprising\n  real mobile a"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2607.28645","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.HC","submitted_at":"2026-05-29T05:48:19Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"80a4133ff665d4c179e5af1af97a1db6f7e4090c7329db3c562ced6c51b23df1","abstract_canon_sha256":"2f5bab7b68bb7e3788e1df77e8dbcc4b890835040ccf8fdea801ec909a90d3a2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-03T00:11:53.633897Z","signature_b64":"BhyTO8zbtZlnLDSBMMsJD+He24mswMO6fdFp3N/251RDimotA5rKPApDuRlrRlFLEP+/Sp98t0Uq+IK1gCrCBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"36b43c31d78c654f5dff3a45c8cac17b6b462e7cd1d3382cc86219411e699239","last_reissued_at":"2026-08-03T00:11:53.631540Z","signature_status":"signed_v1","first_computed_at":"2026-08-03T00:11:53.631540Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Looks Right, Works Right: A Project-Level Benchmark for Multi-Screen Mobile App Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.HC","authors_text":"Cuiyun Gao, Fan Wu, Qing Liao, Yang Xiao, Yiming Huang, Yujia Chen","submitted_at":"2026-05-29T05:48:19Z","abstract_excerpt":"Recent multimodal large language models can convert visual designs directly into executable code, but real mobile products require multiple screenshots to become a buildable codebase with shared components and working navigation. This project-level setting exposes three limits of existing design-to-code benchmarks: they focus on single-page generation rather than complete codebases, cannot evaluate cross-page navigation, and do not measure project-wide maintainability. We introduce MobileForge, the first benchmark for project-level multi-screen mobile app generation, comprising\n  real mobile a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.28645","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2607.28645/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2607.28645","created_at":"2026-08-03T00:11:53.633114+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.28645v1","created_at":"2026-08-03T00:11:53.633114+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.28645","created_at":"2026-08-03T00:11:53.633114+00:00"},{"alias_kind":"pith_short_12","alias_value":"G22DYMOXRRSU","created_at":"2026-08-03T00:11:53.633114+00:00"},{"alias_kind":"pith_short_16","alias_value":"G22DYMOXRRSU6XP7","created_at":"2026-08-03T00:11:53.633114+00:00"},{"alias_kind":"pith_short_8","alias_value":"G22DYMOX","created_at":"2026-08-03T00:11:53.633114+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/G22DYMOXRRSU6XP7HJC4RSWBPN","json":"https://pith.science/pith/G22DYMOXRRSU6XP7HJC4RSWBPN.json","graph_json":"https://pith.science/api/pith-number/G22DYMOXRRSU6XP7HJC4RSWBPN/graph.json","events_json":"https://pith.science/api/pith-number/G22DYMOXRRSU6XP7HJC4RSWBPN/events.json","paper":"https://pith.science/paper/G22DYMOX"},"agent_actions":{"view_html":"https://pith.science/pith/G22DYMOXRRSU6XP7HJC4RSWBPN","download_json":"https://pith.science/pith/G22DYMOXRRSU6XP7HJC4RSWBPN.json","view_paper":"https://pith.science/paper/G22DYMOX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.28645&json=true","fetch_graph":"https://pith.science/api/pith-number/G22DYMOXRRSU6XP7HJC4RSWBPN/graph.json","fetch_events":"https://pith.science/api/pith-number/G22DYMOXRRSU6XP7HJC4RSWBPN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/G22DYMOXRRSU6XP7HJC4RSWBPN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/G22DYMOXRRSU6XP7HJC4RSWBPN/action/storage_attestation","attest_author":"https://pith.science/pith/G22DYMOXRRSU6XP7HJC4RSWBPN/action/author_attestation","sign_citation":"https://pith.science/pith/G22DYMOXRRSU6XP7HJC4RSWBPN/action/citation_signature","submit_replication":"https://pith.science/pith/G22DYMOXRRSU6XP7HJC4RSWBPN/action/replication_record"}},"created_at":"2026-08-03T00:11:53.633114+00:00","updated_at":"2026-08-03T00:11:53.633114+00:00"}