This release makes LifeJet easier to trust before launch: more auth, onboarding, chat, memory, upload, journey, and background failures now report to Sentry with safe context, while the chat surface and memory writes are less likely to fail silently.
Shipped commits
- fix(observability): capture LifeJet route failures in Sentry - Auth callbacks, onboarding, chat, memory, profile, threads, uploads, journey check-ins, artifacts, background jobs, and voice completion now report unexpected failures with route-aware context instead of only falling back to logs.
- fix(observability): repair trace flushes and report telemetry persistence failures - Conversation trace events now store schema-safe error payloads, and trace root/update/flush failures are visible in monitoring when telemetry persistence breaks.
- chore(observability): add Sentry issue readback commands - The release workflow can now verify Sentry API access and list unresolved production issues from the command line without exposing tokens in tracked files.
- chore(chat): add LifeJet chat observability - Chat generation now records more of the pipeline around model messages, context loading, and post-response work so production failures are easier to isolate.
- fix(lifejet): harden memory writes and pattern confidence - Memory and pattern persistence paths now degrade more carefully when optional writes fail, reducing the chance that background state issues block the user experience.
- fix(chat): avoid request cookies in stream finalizers - Streaming finalizers no longer depend on request-scoped cookies after the response begins, which makes post-response chat work safer in production.
- fix(chat): remove viewport slack anchoring - The conversation surface no longer applies extra viewport slack that could destabilize the chat layout.
- feat(marketing): link changelog from footer - Visitors can reach the release history directly from the public footer.
- chore(worktree): harden local worktree setup - Automation and hotfix worktrees now link the local runtime files they need before tests, builds, or deploys run.
Why it matters
When users hit a blocking bug, the team should be able to see where it failed without asking the user to reproduce it. This release moves LifeJet toward that operating model: safer error capture, cleaner production readback, and stronger release checks before tomorrow's users log in.