Skip to content

Releases: Gooseware/opencode-antigravity-autopilot

v2.3.0 - Highest Availability Selection & Quota Enhancements

30 Jan 07:06

Choose a tag to compare

Added

  • Highest Availability Selection - Improved model rotation to select the candidate with the most remaining quota fraction instead of just the first available.
  • Enhanced Quota Cache - Added refreshDate, quotaUsed, and human-readable summaries to the quota cache for better visibility in UIs.
  • Manual Selection Support - Added quota_use_model and quota_auto_mode tools to allow users to lock a specific model while preserving the ability to auto-return to preferred models later.
  • Auto-Return Logic - System now intelligently returns to the user's preferred model once its quota has refreshed, if not in manual override mode.

Fixed

  • Model Name Translation - Refined patterns for Gemini 2.5, Gemini 1.5, Claude 3.5, and more to ensure consistent display names.

v2.2.10

25 Jan 10:50

Choose a tag to compare

Full Changelog: v2.2.9...v2.2.10

v2.2.9

25 Jan 03:04

Choose a tag to compare

Full Changelog: v2.2.8...v2.2.9

v2.2.8 - Model Switch Preference Fix

24 Jan 18:59

Choose a tag to compare

Fixed

  • Model Selector Order - Corrected ModelSelector to strictly respect the user's preference order defined in quota.json. Previously, it was incorrectly picking the model with the highest remaining quota percentage instead of following the priority list.
  • Provider ID Support - Improved providerID detection for more models (including GPT/OpenAI) and added support for explicit provider/model format in model IDs.

v2.2.7

24 Jan 18:12

Choose a tag to compare

Fixed

  • Bumped version for release after manual README fixes

v2.2.6

24 Jan 18:12

Choose a tag to compare

Fixed

  • README in npmjs - Resolved issues with README formatting and visibility on npmjs.com

v2.2.5: Fix Toast Undefined Crash

23 Jan 14:46

Choose a tag to compare

v2.2.4: Fix Toast Crash

23 Jan 14:41

Choose a tag to compare

v2.2.3

23 Jan 14:22

Choose a tag to compare

Fixes

  • quota: properly respect configurable quota threshold in agent integration (fixed hardcoded 0.2 threshold)
  • ui: implement model name translation layer for user-friendly toaster notifications
  • core: expose quota threshold via QuotaManager for integration consistency

Improvements

  • Added model-name-translator utility for mapping API model names to display names
  • Updated toaster severity levels based on configurable threshold

v2.2.2: Documentation Refinement

23 Jan 05:30

Choose a tag to compare