Skip to content

ForgeMind v0.2.1-alpha — 4,000+ Model Aggregator Support & Live Portal

Pre-release
Pre-release

Choose a tag to compare

@DIIIUA DIIIUA released this 14 Nov 14:36
· 9 commits to main since this release

ForgeMind �0.2.1-alpha

🐛 Critical Bug Fix: 1,000–4,000+ Model Aggregator Catalogs (Chat & Settings)

  • Eliminated Duplicate-ID Crashes: Fixed IllegalArgumentException: Key "..." was already used in Compose LazyColumn when opening model pickers for OpenAI-compatible aggregator gateways returning >1,000 models or duplicate entries (such as repeated �imlapi/dynamic-router IDs) in both Chat (ModelPickerBottomSheet) and Settings (ModelPickerDialog).
  • Category Chip Overflow Protection: Replaced uncapped horizontal chip rows with virtualized LazyRow category/FREE chips capped at the top 20 categories by frequency.
  • Zero-Lag 4,000+ Model Search: Added PreparedModelCatalog with O(1) ID maps, pre-indexed search blobs, LazyColumn contentType reuse, filtered/total count badges, capped saved-model chip previews (24 + expand), and skipped redundant local context fallback HTTP probes when /v1/models exceeds 250 models.
  • Automated Regression Suite: Added LargeModelCatalogTests covering duplicate-ID merging, whitespace/blank normalization, category capping, favorite sorting, and <250ms filtering across 3,500–4,000 models.

🌐 New Bilingual Portal & Architecture Docs

Full Changelog: v0.2.0-alpha...v0.2.1-alpha