Changelog

OpenAI costs that hold up, and a Diagnostics tab that tells you why: when OpenAI sends a bad answer, the card says what is missing instead of failing. In detail →

Fixed

• OpenAI card “no JSON object” (a customer wrote in about it): costs now load in sections of up to 31 days, newest first. A section that fails is split until the bad day is found, so one broken answer no longer fails the card.
• Days the app cannot read are named — never counted as zero, never added to a sum. They are retried in the background, and the cost history survives a restart.
• Nothing is quietly wrong: “$12.40 · 2 days missing, 118 catching up”; an unknown “today” reads “Today not available”. The budget brake says “incomplete – no verdict” instead of “normal”, and the daily cost alert stays silent without a reliable “today”. Report, month summary, share image, widget and status line say the same.
• Cost answers with extreme numbers are read: such values are set aside one by one. Amounts are checked for every provider — an unreadable amount shows as an error, not as a silent zero. For xAI an unreadable spending cap shows as “Spending cap unreadable”.
• Grok (xAI): “Team ID (manual)” now works. The app saved the field but never used it; a management key without the right to read the team list failed with a 403 even with the right ID.
• ⌘, opens the real Settings window instead of an empty one named “AI-Cockpit Settings”. Demo mode keeps its example sessions and no longer shows real Claude Code sessions.

New

• Diagnostics tab, rebuilt: one sentence on whether everything is fine, a status light per provider (last success, next request, rate limit), and the last 50 requests of each — path, period, status, duration, size, and for failures the reason and the request ID. Never keys, tokens, names or answer contents. Kept on this Mac.
• “Test now” runs the real request of a card and shows the steps live. A report for support to copy, share or open as a mail draft — the app never sends it. No keys, no account IDs; amounts only on request.

Changed

• A provider that was never asked shows “Not requested yet”, not green; without a success for 24 hours it turns red. “200 · unreadable” shows in the warning color; an expected 404 shows grey. The status line file marks incomplete OpenAI costs.
• Header buttons are back in one row: collapse/expand, pause, usage by project, Your AI month and share as image, each with a tooltip. The “…” menu from 7.6 is gone.

Clear and calm: every card fits its text, outages show once, and nothing fails quietly any more. In detail →

Changed

• Card header in two rows: name and plan badge on top, age and notes below when space runs out. Nothing gets cut off any more.
• One outage bar instead of three copies: a single line under the summary names the provider, the kind of outage and every card it affects, with a link to the status page. Cards keep only the warning sign. Outage texts are our own, in English and German; Copilot (GitHub) joins.
• Calmer cards: bell and ⓘ appear when you point at a window; one forecast line instead of three (“Lasts until the reset · today’s share 14%”).
• Header menu: Refresh and Settings stay, the rest moved into a “…” menu with labels.
• Sparklines with a time axis across the whole window, with a dashed plan line from 0 to 100%.
• One number format: “63%” in English, “63 %” in German; amounts as “USD 12.40”. Today’s share reads “Today’s share: 14% of the week, 0% used”.
• Settings: new pages General and Costs; search jumps to the section and highlights it; limit and credit save at once; ⌘W closes, Return no longer does.
• Status colors meet 4.5:1 contrast in every theme. Tiles carry your own account names; plan badges show only known plans. Widgets: source name on the ring tile, no overflow, age on the control.

Fixed

• Folder access to .claude no longer vanishes after one failed attempt; the app keeps it and asks to renew it. The folder dialog starts in your home folder.
• The sessions card said “no active session” when access was missing — now it says “Access missing”. After a reinstall a hint shows how to get project usage back.
• History survives a reinstall: it now lives in the app group container and moves there once. Settings › History can export and import it as one file.
• OpenAI daily costs are grouped by UTC: the first of the month is back, no false spike alarm. The budget warning works even when one cost card keeps failing.
• A write error could count usage twice; a broken Codex value could crash the app. VoiceOver reads the whole card header and every control in Settings.

Your money and your time in hand: a note before a plan renews, a bell for every reset, a daily share for the week, quiet hours — and what AI cost you last month. In detail →

New

• Renewal calendar: a renewal date and monthly or yearly for Claude, Claude 2 and ChatGPT/Codex (Settings › Display › Subscription costs). A notification 1–7 days before (3 by default) at 9 AM, for Claude with the plan check; from 7 days before the card says “Renews in 5 days (14 Oct)”.
• Reset alarm: a bell in the title row of every window with a reset — also model windows, Codex and Kimi Coding. One notification at the reset, or at once on an early reset; a second click removes it. It comes through quiet hours and pauses.
• Daily share for weekly windows: “9 % left for today so the week lasts” or “4 % over today’s share” — on the cards and in the large widget. On by default (Settings › Display › Quotas).
• Quiet hours (Settings › Notifications, 22:00–07:00, off by default; cost alerts and budget may pass) and pause until tomorrow with the moon button in the panel or the menu bar icon’s right-click menu.
• “Your AI month”: on the 1st, what the month before cost — plans and API, each service, the month before, weekly use and the value of your Claude use. Kept on your Mac for 13 months; share it as an image without account names.

Changed

• Notifications also appear while the app is in front.

Fixed

• OpenAI costs are back. OpenAI now returns costs for the past year only and turns older queries down with HTTP 400; the app asked for three years, so the card stayed empty. Now it reads 364 days and, if OpenAI names a later date, asks once more from there. A refusal now shows the service’s reason too.
• After a reinstall the card of the second Claude account was missing although it stayed signed in. A new switch, “Show second Claude account”, sits in “Sections in the window”, and the card comes back by itself at launch.

For developers, and for clarity: your limits in the Claude Code status line, a note when Claude is done, and a plain answer to “where do I stand?”. In detail →

New

• Your limits in the Claude Code status line (off by default): AI Cockpit writes a small stand.json — percentages, resets, amounts, no keys, no account names — and ships a script for a line like “Claude 63 % · Codex 49 % · OpenAI $2.87”. Settings › Diagnostics copies the command.
• “Claude is done”: a notification when a Claude Code session has finished, next to “waiting for you”.
• A tip on where to switch: at 90 % the card names an account or service with room left — “Codex has 70 % left (5 h)”.
• A plan check from your weekly use (Claude): average and peak of the last four weeks against your plan. An estimate.
• One sentence on where you stand at the top of the panel and in the menu bar tooltip, and (i) buttons that explain the terms.

No surprises: alerts for off-schedule resets and cost spikes, a check on whether your Claude plan pays off, a control for Control Center — and a new look. In detail →

New

• An alert when a window resets off schedule — Codex, for one, sometimes resets after an outage. It waits until the next fetch confirms. On by default.
• A cost alert: credit at OpenRouter, DeepSeek or Kimi dropping fast within an hour, or OpenAI and Anthropic API costs at least twice a typical day. Amount and factor are yours to set.
• “Is your subscription worth it?” The Usage window sets the API value of your Claude Code use against your plan price and says when a cheaper tier would have done. An estimate at API list prices, Claude only.
• A control for Control Center (macOS 26 or later): the tightest value, and a click fetches everything.
• Share your status as an image, Claude account names hidden.
• Three colour schemes — Graphite, Deep Blue, Classic — and a colour per card. Graphite is the new default; Classic brings back the old look under Display › Appearance.

Figures that are honest about their age, a forecast without history and fewer false “signed out” messages. 7.2 takes the place of 7.1 and carries everything from it, from 7.0 and from 6.9 — all three were withdrawn before release and never came out on their own. In detail →

New in 7.2

• A forecast without history under every quota window: “At this pace full at 16:40 — before the reset” or “lasts until the reset (≈ 70 %)”, from the time marker alone. It steps in after launch or with history off, and warns by notification too.
• Wakes up with your Mac: after sleep it refreshes as soon as the network is back.
• Right-click menu on the menu bar icon: Refresh, Settings, Usage, Quit. The second Claude account in the menu bar and the key figures bar, which also offers Kimi, OpenRouter, Grok, Copilot and DeepSeek now.
• GitHub outages on the Copilot card (www.githubstatus.com), like Claude, OpenAI and Kimi. Clicking the widget opens the panel.

Changed in 7.2

• 429, server errors and timeouts no longer mean “signed out”. Figures stay on the card; a rejected key rests for an hour.
• Old figures look old (“⚠ 2 h ago”); a past reset fetches anew. The widget uses the app’s warning levels and thresholds and shows the age per row.
• Anthropic costs fetch only the days not yet stored instead of three years.

From 7.1

• Accounts, rebuilt as an overview. One row per service with its coloured letter as on the cards, the name in large type, a line on what it brings and a status pill — “active”, the last four characters of a stored key, “expired” or “error” in orange, “missing” in grey. Each row opens its own page: sign-in or key, replace and remove, where to get it with a link to the provider, and the usual pitfalls as a highlighted note. A settings search that matches exactly one service opens its page directly.
• Widget: weekly reset in short form — “↻ Thu 4:13 PM”, from a week out just the date — instead of “Reset” and a date that got cut off.

From 7.0

• DeepSeek, the ninth service: your balance, with what you topped up and what was granted. More than one currency shows as separate lines, never added up. When the balance runs out, the card says so. API key under Accounts; the card fetches every 5 minutes and only reads.
• Search the settings. A search field at the top of the settings sidebar finds every switch in English and German, ignoring case and accents. Enter jumps to the first match, Esc clears the field, ⌘F puts you back in it.

From 6.9

• Used or left. Every quota as used (default) or as what is left — on cards, bars, key figures, the menu bar and the widget. Warning colours still follow usage.
• Time marker in the bar: “15 % over plan”, “on plan” or “under plan”, with the time until reset (“in 1 h 42 min”); in the widget too, with a live countdown.
• The widget can be edited — one source or all, used or left — and shows Codex credits. Add the widget once more after updating.
• Calmer refresh intervals (Claude and ChatGPT/Codex every 3 minutes, the rest every 5), slower after repeated throttling, and a random offset so the cards no longer all ask at once.

Fixed

• The “reset” notification never fired; windows and links opened behind the floating panel; a made-up “Opus” weekly line could appear next to “Fable”.
• The widget showed GitHub Copilot with the letter of another service; it now uses “P”, as the app does.
• The Usage window follows a switch between light and dark while it is open.

Where your quota went — and a set of numbers that were wrong.

New

• The Usage window. See where your Claude and Codex quota went — by project, model, effort and hour, up to 365 days back. Raw tokens and the equivalent at API list prices (not an invoice), the estimated share of the 5-hour and 7-day windows, and a weekday × hour heatmap. Worktrees count toward their parent repository. Project names can be anonymised.
• Create analysis saves a self-contained HTML report; CSV export too.
• The app reads back what Claude Code still keeps (by default 30 days) and from then on saves the totals itself — numbers only, stored locally.
• History keeps 180 days by default, selectable 30, 90, 180 or 365; a change takes effect without a restart.
• Monthly budget for all API costs, with a warning at 75 % and 100 %.
• Daily summary as a notification at a time you pick (off by default).

Fixed

• Anthropic API costs showed a hundred times too high. Anthropic reports the amount in cents; the app counted dollars. Old history in cents is converted once.
• Session tokens were counted about 2.7 times. Claude Code writes each answer several times (streaming, branches). Answers are now counted once per message ID.
• Anthropic and OpenAI "today" and "this month" follow UTC.
• The Kimi balance shows the currency of your platform (yuan on the China platform). Copilot labels AI credits as AI credits, also on the history tile, and counts in the monthly total.

Changed

• The app is now built on macOS 15 and 26; it still runs from macOS 14.

Fixed

• The ChatGPT card showed the Codex credit balance as a dollar amount. It is not one. OpenAI keeps usage credits and the dollar wallet apart, and credits have no cash value. Balance and the amount spent since the start of the month now read as credits, with no conversion to dollars. Demo mode shows a plausible credit count to match.

Changed

• The App Store listing is new: screenshots from the running app with all eight services, a new subtitle and new keywords.

Fixed

• GitHub Copilot showed nothing on converted accounts. On 1 June 2026 GitHub moved personally paid plans off premium requests and onto AI credits. The card only ever asked for premium requests, so anyone already converted saw an empty card and no error explaining why. It now asks for both and shows whichever one your plan actually bills.
• The Claude fallback model still pointed at Opus where it should have pointed at Fable.

Nothing else changed in this release. Both fixes had already landed in the shared core that the iPhone app builds against — they were simply missing from a Mac build.

New

• GitHub Copilot, an eighth provider, on both Mac and iPhone. It only reports figures for a personally paid Copilot plan — GitHub exposes usage for company- or organization-managed seats at the organization level only, so those accounts show nothing on this card.
• ChatGPT numbers now come from a live fetch after a one-time sign-in, instead of only reading whatever Codex had already logged locally on this Mac.
• A balance figure on the ChatGPT card, and an outage indicator on any card whose provider is down.
• Desktop widgets, a compact menu bar capsule showing up to four values at once, a segment switcher above the cards, and a freely arranged key-figures bar.
• A tip jar under Settings → About — voluntary, unlocks nothing, changes nothing.

Fixed

• The desktop widgets froze over time. They asked to be redrawn thousands of times a day and ran into the system limit, which throttles them. This is the reason 6.4 was withdrawn before review.
• Demo mode kept its sample figures for barely a minute before the refresh timers overwrote them — and it showed some real account values instead of samples.
• A card whose sign-in had broken offered only a Retry that could never help. ChatGPT, Kimi and Copilot now point at the place where the problem can actually be fixed.
• The key-figures bar dropped a tile whenever its value was missing and rearranged itself. It now keeps the last known value in place, greyed out.
• The connections list was incomplete, in the app and on this site. Cancelled requests no longer count as a provider outage.

A planned 6.2 was folded into this release — its two fixes ship here.

New

• One line that answers «what does AI cost me this month, all in?» Enter your subscription prices once (Settings → Display → Subscription costs) and the cockpit adds your live API spend on top. Kimi and Grok stay out of the sum — they only report a balance, and an honest number beats a complete-looking one.
• Feedback and rating, side by side. Two equal links in Settings → About; and the app may now ask for a rating itself — rarely, only after days of successful use, never after an error.

Fixed

• The plan badge on the Claude card. It was never fetched: this client's tokens carry no plan, and the badge came solely from ~/.claude.json, the local Claude Code store. In the Mac App Store build that file is out of reach without a folder grant, and a second Claude account would have been given the first account's plan — that file knows only one. Now read from api.anthropic.com/api/oauth/profile, one field (organization.rate_limit_tier) per token, so per account, then kept in the keychain. Shipped first on iPhone as 2.2.2.
• The setup assistant appeared in German for everyone. Every one of its strings was missing from the English catalogue, so each fell back to its German key. English is this app's default language regardless of the system, which made the feature 6.1 leads with unreadable for most people meeting it. The iPhone edition was never affected — its catalogue is extracted at build time.
• Ten more catalogue gaps closed in both languages — among them the second-account description and the demo-mode section — and a new test now checks the source code against both catalogues, so this class of bug cannot slip through again.

New

• A second Claude account. Settings › Accounts now offers a second one, with its own card, its own mark ("C2") and its own colour, and both accounts can be named. Signing in a second time used to silently overwrite the first.
• A demo mode. Settings › Accounts › Try it out fills every card with sample figures — no sign-in, no keys. The iPhone edition has had one since 1.0; the Mac never did. Nothing is fetched and nothing overwritten while it runs.
• A setup assistant on first launch. Four steps walk you through what the app does, which services you use and the credentials for them. Skip works on every step.
• "Reset everything". Clears both Claude sign-ins and all API keys from the keychain, resets every setting, signs out of launch-at-login and the hotkey, deletes the recorded history and releases granted folders. Keys stay valid at the services themselves.

Changed

• The Anthropic API card has its own colour (an A mark instead of sharing Claude's orange C).
• Claude is a shade lighter in dark appearance, to leave contrast room for the second account's card.
• The Grok card no longer looks disabled — it had shared its lightness with muted secondary text.

Missing from the source changelog file, though it shipped and is live in the Mac App Store — summarised here from the site's own release notes.

• Two more providers: OpenRouter, where one account reaches hundreds of models, and Grok from xAI. Both show what you have spent and what is left.
• Cards can be dragged where you want them, within a column or across to the other one; the arrangement is remembered, with a reset in Settings.
• The menu bar now shows what you pick — two settings choose the first and second figure from every window the app currently has.
• The Anthropic API cost card was broken and is fixed: it asked the server for the wrong address and blamed your account for the empty answer. A rejected key now says plainly what it needs.
• The light theme is properly readable: cards sit lighter than the ground instead of darker.

Each card now carries its own coloured letter mark — C for Claude, G for ChatGPT, O for the OpenAI API, A for the Anthropic API, K for Kimi, S for the sessions — matching what the iPhone edition already showed.

• A cancelled fetch is no longer reported as a failure; a card keeps whatever figures it already had.
• Requests now have a real ceiling — two minutes — instead of relying on a system default that could hold a connection open for days.
• No more raw system error text on a card; readable wording instead (no internet connection, service unreachable, secure connection failed).

• Settings moved from tabs to a sidebar, after a macOS 27 beta moved the tabs of a TabView into the title bar and collapsed whatever did not fit.
• Regression guards added for sidebar width and label wrapping.
• About now names the developer: Albert Frick.

• Settings window fixed: no longer a hard-coded 780×780 sheet, now its own resizable window with a sensible minimum.
• A second, independent cause of the same squeeze removed — the panel behind the sheet no longer resizes to the dashboard's shrinking width.
• Regression guard added; CI now runs on macOS 14, 15 and 26.

• Appearance: selectable dark (default), light or system.
• Language: selectable, English default regardless of system language.
• Mac App Store build: fully sandboxed, with security-scoped folder grants for ~/.claude and ~/.codex. Released on the Mac App Store on 18 August 2026 as 5.1 (56).
• Localization fix for ten format strings and seven core error messages that silently fell back to German.
• License: the repository is private and proprietary from here on; versions up to 5.1 remain available under GPL-3.0 to those who obtained them while the project was open source.

First public release.

• Bilingual: full German and English interface (316 localized strings).
• Renamed to AI Cockpit, new app icon, new bundle identifier com.ipstyle.aicockpit.
• OpenAI credit balance: derived from a recorded top-up amount minus billed daily costs, chosen deliberately over storing a browser session key.
• Security review 4.2 (Claude Fable 5): no critical finding; two important ones fixed.

OpenAI costs that hold up, and a Diagnostics page that tells you why: the card shows what it knows and says what is missing, instead of “No JSON object”. In detail →

Fixed

• OpenAI card “No JSON object” (a customer wrote in about it): the reason is named — empty, not an object, cut off, or a parser error with its position. One broken day no longer empties the card.
• Costs load in sections of 31 days; an unreadable section is halved until the bad day is found. Days that were read stay in the App Group storage that app and widget share. In steady use: one request instead of seven.
• The gap sits next to the figure: “USD 12.40 · 2 days missing, 118 catching up”. An unknown “today” reads “not available”, never 0. Cost brake and alert do not judge on a gap: “incomplete – no verdict” instead of “normal”.
• Widget, watch list and card show the gap sign; the watch never starts a cost run. Your old history moves to the new storage before it is deleted.

New

• Diagnostics (Settings › Accounts › Diagnostics): a light per provider, a timeline of the last requests, “Test now” with live steps, and a report to pass on. The report blacks out IDs, shows amounts only on request and never leaves by itself.
• Request log in the app, the widget and the background run: path, period, status, duration, size and reason — never keys, never the text of an answer. Your AI month and the monthly notice name gaps: “API incomplete – 2 days missing”.
• Grok (xAI): “Team ID (manual)”, a new field in the Grok settings for management keys that may not read the team list.

Clear and calm: every card fits its text at any text size, outages show once, and nothing fails quietly any more. In detail →

Changed

• Card header stays readable at every text size; the summary moves below the name when needed.
• One outage bar above the cards: provider, kind of outage and every card it affects. Tap for the status page. Cards keep only the warning sign.
• Header: “…” menu, Refresh, Settings. Order, Fold all, Share and Your AI month sit in the menu.
• Calmer cards: tips under “More”, reset alarm in the context menu of a window row.
• iPad: independent columns — two from 700 pt, three from 1100 pt.
• Cost cards with a 30-day line and a spike note. One format: “63%”, “USD 12.40”. Status colors meet 4.5:1 contrast; plan badges only for known plans.

Fixed

• One failed fetch no longer empties anything. On 429, a server error or without network, card, widget, watch and Live Activity keep their last figures, marked with their age.
• Watch: answers with Claude and ChatGPT within 7 s; an older state no longer replaces a newer one; “iPhone not reachable” without a connection.
• Reset alarm no longer fires from stale data; no double “reset” notice. Lock Screen shows the number first and tells “Claude” and “Claude 2” apart.
• Live Activity and complications show their age; outages reach tiles and the watch. VoiceOver reads the whole card header.

Your money and your time in hand: a renewal date for each plan, a reset alarm, a daily share, quiet hours, “Your AI month” — and six Shortcuts instead of two. In detail →

New

• Renewal date and monthly or yearly per plan (Settings › Subscription costs › Renewal), a notification 1–7 days before at 9 AM, and from 7 days before a line on the card.
• Reset alarm: a bell to the right of each window (Claude, Claude 2, ChatGPT/Codex). Planned on the device, so it comes on time with the app closed; also in quiet hours. Not on the Apple Watch.
• Daily share on the Claude and ChatGPT cards and in the large widget: “9 % left for today so the week lasts”. On by default (Settings › Appearance).
• Quiet hours (Settings › Notifications, off by default) and pause until tomorrow — in the settings, as a Shortcut, with a band above the cards.
• “Your AI month”: on the 1st, what the month before cost. Tap the total line “AI this month” or open aicockpit://monat; share it as an image without account names.
• Six Shortcuts: new are status of all services, daily share, set reset alarm and pause notifications until tomorrow. “New in 2.16” in the app.

Fixed

• OpenAI costs are back. OpenAI now returns costs for the past year only and turns older queries down with HTTP 400; the app asked for three years, so the card stayed empty. Now it reads 364 days and, if OpenAI names a later date, asks once more from there. A refusal now shows the service’s reason too.

Clear at a glance: one sentence on where you stand, a tip on where there is still room, and widgets that open the right card. In detail →

New

• One sentence on where you stand — “All in the green”, “Claude is getting tight today” — at the top of the app, in the large widget and on the Apple Watch.
• A switch tip: when a window is 90 % full and the other service or your second Claude account has more than half left, the card says so.
• A plan check from your weekly peaks (Claude): would a smaller tier do? An estimate, a verdict after two full weeks.
• (i) buttons that explain the windows, credits and alerts.
• Widgets open the matching card, the Live Activity the tightest one, and “New in this version” sits in the settings.

No surprises: alerts for off-schedule resets and cost spikes, a Live Activity with up to four bars, a control and a new look. In detail →

New

• An alert when a window resets off schedule (on by default) and a cost alert for credit at OpenRouter, DeepSeek and Kimi and for OpenAI and Anthropic API costs (off by default). Both come when the app fetches; in the background iOS decides how often.
• A Live Activity with up to four bars — by default the two tightest windows, or the ones you pick — on the Lock Screen and in the Dynamic Island.
• A control for Control Center, the Lock Screen and the Action button (iOS 18 or later), on the Apple Watch from watchOS 26.
• Higher in the Smart Stack when a window is nearly full or about to reset — on the iPhone, and on the watch from watchOS 26.
• Share your status as an image, account names hidden.
• Three colour schemes — Graphite, Deep Blue, Classic — and a colour per card, carried into widgets, watch and Live Activity. Graphite is the new default; Classic brings back the old look under Settings › Appearance.

When will it run out? A forecast under every window, a refresh button in the widget, and fewer false “sign in again” messages. In detail →

New

• A forecast under every window: “At this pace full at 6:48 PM — before the reset”, or that it lasts until the reset. From the time marker alone, no history needed. Also in the large widget, on the rectangular Lock Screen widget and in the rectangular watch complication — and, if you want, as a notification before a 5-hour window fills.
• A refresh button in the Home Screen widget: one tap fetches every service, without opening the app.
• A widget above the clock (one line on the Lock Screen), and a value you pick on the small widget — 5 hours, week or credits.
• GitHub outages on the Copilot card, from www.githubstatus.com.

Changed

• “Sign in again” only when the service really turned the sign-in down; throttling, 403 and server errors pass by themselves.
• A lean background refresh: only Claude and ChatGPT, within 20 seconds. The Apple Watch asks for fresh figures when you open it or raise your wrist.
• “Reset everything” also clears widget, watch and Live Activity. Demo mode shows no real accounts.

Widgets that stay fresh overnight, and a new Accounts page. In detail →

New

• The widget renews an expired sign-in itself — Claude, the second Claude account and ChatGPT/Codex. Up to 2.11 only the open app did, so after a few hours the tile said “open” and stood still. App and widget agree first, so a sign-in is never renewed twice.
• Background refresh: iOS may wake the app now and then to fetch the figures for the widgets. How often is up to iOS.
• A new Accounts page: every service with its coloured mark, the name in large type and its state at a glance — active, missing or the last four characters of the key — and a page per service with a link to where the key comes from and the pitfalls as a visible note. ChatGPT/Codex can be signed in and out there.

Changed

• Lock Screen, weekly window: “↻ Thu 4:13 PM” (from a week out just the date) instead of a cut-off “Reset 4.…”. The age of a figure shows there only from 15 minutes; on the Home Screen always.

Fixed

• Lock Screen widgets that stayed on “open” after a few hours — the widget now renews the expired sign-in itself.
• If the provider turns a renewal down, the tile says “sign in again” instead of “open”. A late renewal never undoes a sign-out.

Widgets still get a daily budget of reloads from iOS; only the countdown runs to the second. 2.12 makes each reload end with fresh figures.

DeepSeek and a watch complication you set up — and everything from 2.10, which was withdrawn before release and never came out on its own. In detail →

New

• DeepSeek, the ninth service: the balance per currency — available, of which topped up and granted; a second currency on its own line, never added up. Also in the widget, demo mode and setup assistant; on the watch in the list of sources.
• A watch complication you set up, next to the old one: source, value (most urgent window, 5 hours, week, credits) and used or remaining. The old complication stays unchanged.

From 2.10

• Used or remaining on cards, widget and Apple Watch; colours follow usage.
• The widget can be set up, on the Lock Screen too, fetches every service itself and shows Codex credits. Add your widgets again after updating.
• Time marker in the bar and a live countdown to the reset on the Lock Screen.
• Calmer refresh intervals, slower after repeated throttling.

Fixed

• The privacy page in the app is more precise about the widget, the tip and what the widget stores; “Sign out and delete everything” now also clears what the widget keeps for its own refresh.

A brake on costs, and a fuller total.

New

• Monthly budget for all API costs (USD), under subscription costs. The total line reads "API $x of $y" and turns orange from 75 % and red from 100 %. With notifications on, each threshold warns once a month. Last month's figures raise no warning on the first of the month.
• GitHub Copilot counts in the monthly total.

Fixed

• Anthropic costs for "today" and "this month" follow UTC, as Anthropic's own report does.

Fixed

• Anthropic API costs showed a hundred times too high. Anthropic reports the cost in cents as a decimal string — "123.45" means $1.23. The app counted it as dollars. It now divides by 100, in one place, and the old figure no longer lingers after the update.
• The Kimi balance always said dollars. The China platform bills in yuan. The card now shows the currency of the platform you signed in to.
• Copilot called AI credits "requests". Since 1 June 2026 GitHub bills personal plans in AI credits. The card now labels the figure by the billing your plan actually uses.
• OpenAI's "today" and "this month" ran on local time, while OpenAI's daily figures run on UTC. Both follow UTC now.

Fixed

• The ChatGPT card showed your Codex credit balance as a dollar amount. It is not one. OpenAI keeps usage credits and your dollar wallet apart, and states that credits have no cash value. The line now says what it is — credits.
• GitHub Copilot now appears in the demo mode as well. Until now the demo showed seven of the eight services.
• One line on the Kimi card was still German in the English app. It now reads "no usage figures via the API".

New screenshots from the running app, and the description now says outright that iOS 17, iPadOS 17 and watchOS 10 are enough.

New

• It runs on iOS 17 now, not iOS 26 — and on the Watch from watchOS 10 instead of watchOS 26. Three calls sat above that line and now sit behind version checks, so nothing was dropped to get there. If the app would not install on your phone before, try again.
• Feedback, the tip jar and the links moved into the settings list, just above Reset everything, instead of hiding under About. Two links are new: Security and the privacy policy.

Fixed

• GitHub Copilot showed nothing on converted accounts — the same fix the Mac gets with 6.6. GitHub moved personally paid plans from premium requests to AI credits on 1 June 2026; the card now asks for both.
• The Kimi footnote left out the coding quota, one comparison still said five providers when there are eight, and a note claimed the Mac app costs money. It is free.
• The privacy manifest declared DiskSpace and SystemBootTime, neither of which appears anywhere in the source. Both are gone.
• A dead bridge card was removed.

A planned 2.4 never reached the App Store — everything it carried ships here.

New

• GitHub Copilot, an eighth provider — the same one the Mac gets with 6.5. It only reports figures for a personally paid Copilot plan; GitHub exposes usage for company- or organization-managed seats at the organization level only.
• An outage indicator on the Claude, OpenAI and Kimi cards, read from those providers' own status pages. Two deliberate differences from the Mac: with no hover text to fall back on, the message stands as its own full line, and it is fetched with the foreground refresh passes — an iPhone runs none in the background.
• A balance figure on the ChatGPT card.
• A tip jar under Settings → About: three amounts (Espresso, Cappuccino, Dinner), repeatable, unlocking nothing and changing nothing, and the hint at the foot of the card list falls silent once you have given. With it, the download is free again — 2.3 carried a price — and the tip jar is the app's only in-app purchase. The Mac's tip jar has been there since 6.5.

Fixed

• The widget and the watch complication froze over time. Both asked to be redrawn on every refresh tick and ran into the system limit, which throttles them — the same cause as on the Mac, with the wrist affected on top.
• A refused sign-in left a dead button. A card whose stored credentials had been rejected offered "Sign in", and pressing it did nothing. All eight cards were affected — and it was the one state that only that button leads out of.
• The Copilot card offered a Retry where the token was missing a permission. Retrying never helps there; it now points at the settings, where the token can be replaced.
• The connections list under Settings → Privacy and Security was wrong: api.github.com was missing, and api.kimi.com was listed although this edition never contacts that host.
• Five Copilot error messages were missing from the English catalogue and appeared in German on English devices.
• The Anthropic card has its own colour, and a card without a provider no longer borrows the accent colour — it was indistinguishable from the Grok card.
• The light theme had drifted away from the Mac's. Small print under the figures reached only 3.84:1 against the card, below the threshold; it is 5.43:1 now.

The first update after the debut, and a large one: 2.0, 2.1, 2.2, 2.2.1 and 2.2.2 were built but never submitted on their own — all of them ship here.

New

• OpenRouter and Grok (xAI), on the phone and on the watch. One OpenRouter account reaches hundreds of models; Grok shows the balance of the xAI account and the spending cap that applies. xAI needs a management key from the console, not the xai-… key used for model calls.
• Refresh rates per source, adjustable, with the same defaults and ranges as the Mac — and the clock now runs while you are looking, not only when the app comes to the front. Foreground only: iOS lets no app fetch in the background, and the settings page says so rather than promising an update that cannot happen.
• A setup assistant on first launch. Four steps instead of seven grey cards full of denials; Skip works on every step.
• A second Claude account with its own card, its own mark ("C2"), its own colour and a name of your choosing. Signing in a second time used to overwrite the first without a word.
• Siri and Shortcuts. "How full is my Claude window?" gets an answer in one sentence, the age of the figures included. The command never goes to the network — it reads the same store the widget draws from. A second command brings the app to the front so it can sit inside a longer shortcut.
• A Live Activity for the window that is pressing. Once the Claude card passes the warning threshold and a reset time is known, the Lock Screen and the Dynamic Island count down to it. Local only, no push token: no server is told how full anyone's subscription is. The price is named honestly — the percentage moves when the app has run, not in between.
• The Mac's monthly-total line. Enter your subscription prices once and the app adds the running API costs on top. Nothing is converted, and the line says so as soon as both parts are involved. Kimi and Grok stay out of the sum — they only report a balance.
• Feedback and rating, side by side under About, with no mood question in front of them; and the app may ask for a rating itself — rarely, and never while a visible card shows an error.

Fixed

• The plan badge on the Claude card was never filled. The stored token carries no plan — it never did, not even right after signing in — and the Mac only shows one because it reads ~/.claude.json, a file no iPhone has. It now comes from api.anthropic.com/api/oauth/profile, one field, asked once per account, so the second account gets its own.
• The accounts row said "signed in · 4 of 3 keys". The denominator was a literal from before OpenRouter and xAI; four further places carried the same stale number, among them the privacy page and its list of network destinations.
• The demo mode showed a projection the app does not have. It is out until the history that would make it real exists.
• The Grok card looked disabled — its colour sat at the same lightness as muted secondary text. And the card header wrapped at Kimi and Grok.

Changed

• The Anthropic API card has its own colour and an A mark; with a second Claude account there would otherwise have been three warm cards in a row. Claude is a shade lighter in dark appearance to make room for it.

The first release. 1.0 and 1.0.1 were submitted and replaced before they reached anyone, so what arrived is 1.1 — with the Apple Watch already in it.

New

• Five cards — Claude, ChatGPT/Codex, the OpenAI API, the Anthropic API and Kimi. Every card fetches its figures on the device itself; no Mac is needed. Cards you do not use can be hidden and the rest reordered, and each source runs for itself: if one breaks off, the others are untouched and whatever a card already had stays standing.
• Widgets in five sizes — small, medium and large for the Home Screen, circular and rectangular for the Lock Screen. One line per service in its own colour and with its own mark, label and figure kept apart, and the age of the numbers alongside.
• Apple Watch. Three levels: a verdict in one sentence, the sources as a list below it, and a ring per window in the detail — paged with the crown rather than scrolled, because scrolling cut the second ring through its own number. Plus a complication in all four shapes, legible in the single-colour rendering because every warning level has its own symbol, not just its own colour.
• The watch shows, it does not fetch. No token, no key and no secret lives on it; finished display values arrive from the iPhone over Apple's own pairing. A refresh token may be redeemed in one place only — a third party in the race would eventually lose and sign you out.
• No empty start. App and widget open with the last known figures and write down how old they are, instead of going blank while fetching.
• A demo mode with a full set of plausible figures and no sign-in at all.
• Notifications at reached limits and starting windows, each one switchable and all of them off to begin with.

Fixed

• Pull to refresh cancelled the fetches it had started. Every source ran in one group whose root was SwiftUI's task, so when SwiftUI discarded it they broke off together and three cards read "cancelled" although no service was down. Each source now runs in a task the app owns.
• The ChatGPT sign-in hung silently: the callback path differs per service (Claude /callback, Codex /auth/callback) and was hard-wired.
• Requests had no real ceiling. timeoutInterval only measures the pause between two signs of life, so a cost fetch could run for 35 minutes and hold its source as "in flight".
• Hidden cards and newly entered keys took effect only after a restart — a global fetch lock swallowed the second pass. Each source now locks for itself.

OpenAI costs that hold up, and a Diagnostics page: the OpenAI fixes and the diagnostics of Mac 7.7 and iPhone 2.18. In detail → · Download

Fixed

• OpenAI card “no JSON object” (a customer wrote in about it): costs load in sections of up to 31 days, the current month first; a section OpenAI answers unreadably is halved until only the bad day is missing. The error line now says what happened — “OpenAI: answer unreadable · costs 18.04.–15.10. · HTTP 200 · 41 KB”.
• Gaps are named, not filled: “12.40 $ · 2 days missing”, “Today not available”. The days live in %APPDATA%\AI-Cockpit\kosten, are fetched later and survive a restart. Cost brake and cost alert give no false all-clear.
• Unreadable amounts are no longer counted as 0 — for Anthropic, GitHub Copilot, Kimi, OpenRouter, xAI and the Codex credit.

New

• Diagnostics page in Settings: one sentence, a light per source, the last 50 requests, “Test now”, the cost coverage of OpenAI, and a report for support to copy or open as a mail. The app never sends it; amounts only on request; IDs of accounts and keys always blacked out.
• Request log (%APPDATA%\AI-Cockpit\diagnose): no headers, keys or answer contents.

Changed

• Fewer requests: project names and key counts are remembered for six hours. Reset deletes costs, project names and the log thoroughly. Dates follow the language of the app. Files are replaced more robustly when a virus scanner holds them open.

The file is not signed.

The features of Mac 7.5: your money and your time in hand. 1.8.1 came out a few hours after 1.8.0 and includes it. In detail → · Download

Fixed in 1.8.1

• OpenAI costs are back. OpenAI now returns costs for the past year only and turns older queries down with HTTP 400; the app asked for three years, so the card stayed empty. Now it reads one year and, if OpenAI names a later date, asks once more from there. A refusal now shows the service’s reason too.

New in 1.8

• Renewal calendar next to the subscription costs, with a notification 1–7 days before at 9 AM and a line on the card in the last week.
• Reset alarm: a bell next to every window with a known reset — Claude for each account, model windows, ChatGPT, Kimi Coding. It survives a restart and comes through quiet hours and pauses.
• Daily share for weekly windows, quiet hours (Settings › Notifications) and pause until tomorrow in the tray icon’s menu.
• “Your AI month” as a window from the tray icon’s menu and as a PNG, with fixed service names instead of account names.