2.0.67-beta.9d63abe

• fix: cache hit rate stuck at 0% + image_generation request counters + auto-precision display (#423)
• feat: track image_generation tokens separately + real-upstream stress test (#422)
• fix: recover from previous_response_not_found instead of passing it through (#418)

2.0.67-beta.da80d30

• feat: persist home account filters (#401)
• feat: surface cache hit rate in dashboard usage stats (#416)

2.0.67-beta.6abf6ae

• fix: recover from previous_response_not_found instead of passing it through (#418)

2.0.67-beta.5f93e05

• fix: recover from previous_response_not_found instead of passing it through (#418)

2.0.67-beta.2c73e13

• feat: surface cache hit rate in dashboard usage stats (#416)

2.0.67-beta.a0f87a9

• fix(ci): force bash shell on Pack steps so Windows runner accepts -continuation (#414)

2.0.66

• feat: add gpt-5.5, gpt-image-2 to catalog; switch default to gpt-5.4 (#408)
• fix: address ollama bridge review followups
• fix: harden ollama bridge review issues
• feat: integrate built-in ollama bridge

2.0.65

• fix chat suffix model routing and stabilize tests

2.0.64

• Localize log pagination and tighten packaging tests
• fix: 修复缓存统计和 WebSearch 转换 (#392)
• fix: sync log controls and preserve zero-latency display
• feat: separate logs settings drawer and sync enabled state
• feat: add persisted LLM-only log capture mode
• fix: correct logs pagination and honor body capture setting
• fix: tighten logs pagination state handling
• feat: add dashboard logs tab
• feat: 本地化 logs 的相关设置项
• refactor: bound request logs and add coverage
• feat: add dashboard logs tab
• fix
• feat: add log tab
• feat: add logs tab and backend
• test: 添加简易测试 test:local-chat 只覆盖了各个模型的 openai 格式的 curl 请求
• test: 添加简易测试 test:local-chat 只覆盖了各个模型的 openai 格式的 curl 请求

2.0.62

• refactor: open tests/ and consolidate test directory structure (#373)