PRISMSAI is now available on the App Store
PRISMSAI 1.0 is publicly available as a free iPhone and iPad app in supported storefronts outside mainland China. Its official Apple ID is 6801998853.
BUILDING IN PUBLIC · SINCE JULY 2026
A crawlable record of architecture, product, knowledge, and live-runtime work.
PRISMSAI 1.0 is publicly available as a free iPhone and iPad app in supported storefronts outside mainland China. Its official Apple ID is 6801998853.
Build 21 passed Apple upload processing and the complete submission—with designed screenshots, review access, privacy disclosures, and public support pages—is now Waiting for Review.
Local date, clock time, IANA time zone, strict-future validation, supplemental input, and the real persisted reminder result now stay in one schedule-agent path.
The signed App archive defines the ABI; Agent packages declare compatible App ABIs, and a read-only release gate checks both before publication.
Signed Agent, Schema, and Skill packages verify and switch atomically, pin one release per task, and keep compatible native builds working.
Public registration is closed. Roster-approved official addresses receive single-use activation and recovery mail without storing raw activation tokens.
Privacy, Support, Terms, Community Guidelines, account deletion, eight privacy-label categories, and designed product screenshots are ready for review.
The foresight agent creates a cue only when authorized profile, memory, knowledge, time, place, weather, health, calendar, or reminder context makes it useful.
A bounded knowledge map, hybrid Top-K retrieval, and on-demand locator recall connect personal and shared sources without overflowing agent context.
English becomes the default public edition, with matching English and Chinese product sections, update history, community entry points, structured data, and feedback flows.
PRISMSAI defines the boundary between on-device agents, the stateless model relay, the isolated control plane, provenance-first memory, and consented iOS system capabilities.
Ask, memory, personal knowledge, agent stages, and task-appropriate text, image, audio, or interactive HTML deliverables move into one persistent iOS workspace.
Business agents retrieve school and external knowledge before source-dependent answers. Unsupported claims remain stable knowledge gaps instead of model guesses.
The private student assistant moves to iOS while the public website focuses on product architecture, dated updates, feedback, and community.
End-to-end student tests begin treating the actual agent chain, evidence consumption, and final artifact—not a plausible Intake response—as completion criteria.
A full refresh checks 517 URLs, saves 502, and keeps web pages, PDFs, documents, and access errors in distinct inspectable states.
Official school ingestion grows to 517 URLs and 592 document pages, updating 286 knowledge objects and 2,407 evidence records.
Agents select a correction skill from the actual stage and failure type, preserve valid work, and rebuild the complete failed object coherently.
Continuous scenarios cover courses, dining, clubs, residence life, communication, and campus services with real-source and human-confirmation requirements.
The first expanded school-site scan fetches 160 URLs, saves 151, and updates 143 knowledge objects with 932 evidence records.
Fifty-eight courses are rebuilt from the official catalog with department, credit, level, recommended grade, offering status, and page-level evidence.