OAuth account experience — 2026-07-21

• Added secure EveryAPI OAuth sign-in with automatic access-token refresh and revocation.
• Added Account & Usage for wallet balance, today and seven-day usage, and top models.
• Ask, model selection, and Recent Requests now share one authenticated API layer with consistent timeout, retry, and sign-in recovery behavior.

Service Status & multi-turn Ask — 2026-07-21

• Ask EveryAPI — upgraded from one-shot Q&A to a persistent multi-turn chat: follow-up questions carry the full conversation as context, streaming can be stopped mid-flight, every reply shows model / token / latency stats (real usage via stream_options.include_usage, falling back to a local estimate), and the model can be switched mid-conversation (⌘M). The conversation survives across command invocations (LocalStorage, capped at 40 messages) — "Continue Last Conversation" resumes it.
• Service Status (new command) — live status of upstream AI providers, aggregated by the EveryAPI gateway from the vendors' public status pages (/api/upstream-status). Shows a severity-sorted provider list with status indicators, a 24-hour history, affected components, and active incidents. The public endpoint works without signing in.

Initial Version — 2026-07-21

First public release.

Commands

• Ask EveryAPI — Form (prompt + model picker) → streamed answer in a Detail view. Works against any model the EveryAPI account exposes.
• Account & Usage — shows the signed-in account, wallet balance, request activity, and model usage, with dashboard actions for deeper account management.
• Switch Default Model — lists every model the account can see (search-and-pick → LocalStorage). The selected default is used by Ask.
• Recent Requests — shows recent calls with model, tokens, cost, latency, and request ID.
• Service Status — shows current provider health, recent availability, and active incidents.

Preferences

• Base URL (default https://api.everyapi.ai/v1)