Drafts persisted the whole NewSessionForm, including `directory`, so the
last path picked through the browse (+) button came back on the next
presentation and seeded the browser. Web's draft deliberately omits the
directory; match it:
- readDraft/writeDraft strip `directory`, and old drafts with a stored path
are ignored on restore.
- A restored draft now seeds its saved machine's default directory (or the
most recent valid spawn path) unless the user edited the field.
- A directory-header "new session" still uses that exact initial path.
- Keep the directory tests in sync and teach the fake HTTP about the Claude
custom/machine model catalogs; update the long-message probe for the
selection-sheet closure signature.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Decode DSH sessions by accepting workspace-write / danger-full-access
permission modes (the missing cases made the whole session detail fail to
decode, so pending ask_user_question requests never reached the card).
- Carry background task snapshots from Session/SessionPatch and let the
composer chip list the running task descriptions.
- Compact the nav subtitle to project + machine, add a tap-through session
info sheet, and hide the duplicated status wording.
- Mute archived rows: desaturated agent icon, secondary title, hint-coloured
detail/timestamp and an archivebox indicator.
- Add a Select Text action backed by a read-only UITextView so arbitrary
ranges (including cross-paragraph) select with native handles; enable
inline table selection too.
- Point home widget card taps at their session instead of the list, and keep
the list link only as the empty fallback.
- Update LastSeen tests for the optimistic read watermark semantics.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Copy a session citation from the list (native parity with the web
reference text) and open the paired session list via
hapicompanion://sessions.
- Load hub-wide custom Claude models in New Session with machine-catalog
fallback, and show the transcript-resolved model in chat chrome.
- Render synthetic Claude API failures as error events instead of leaking
the <synthetic> model or zero usage into the context meter.
- Refresh the session list returning from chat, request the list from Home,
and update the home widget surfaces.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Fetch Claude models from the machine catalog, expose the full grouped picker with per-model efforts, and remember launch defaults.
- Add directory quick-create, known-directory suggestions, live machine CPU/memory in filters, and the DeepSeek agent icon.
- Add the collapsible session status card (goal, tasks, subagents, terminals, team), context-usage meter, and compaction/background/fast/collaboration composer chips without the redundant inline model controls.
- Add code block line numbers, wrapping and collapse; per-message share and rewind; a floating session search that does not push the list; and move sign out to the bottom of Settings.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Cut the widget reload coalescing window to 15 seconds, the hub silent-push throttle to 30 seconds, and the timeline fallback to five minutes so active sessions stay closer to the app.
- Restyle the Live Activity lock screen and Dynamic Island with the app icon, aligned time/status columns, task/result detail, fixed spinner and timer widths, and explicit edge padding.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Add a live session status line and project/model/effort metadata to the chat header, plus a compact composer strip for model, effort, permission and connection state.
- Add per-message timestamps, copy, and duration/model/token details; switch user bubbles to a neutral surface, card the reasoning block, and timestamp event rows.
- Surface tool start/duration in summary cards and give list agent icons a native surface tile while preserving grouping and row metrics.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Include the latest assistant preview in the silent-push fingerprint so result-only changes wake the phone; the status-only fingerprint missed results that landed while a session stayed working.
- Coalesce WidgetKit reloads (60s) and reload only installed widget kinds, with a fallback when configuration lookup is empty; the previous every-update/all-kind reloads exhausted the system redraw budget.
- Wait for AppModel/hub restoration on a cold silent-push launch instead of clearing the last good snapshot.
- Use a 15-minute widget timeline fallback and 60-second hub push throttle.
- Add the medium Session Detail layout with in-card paging, plain-text previews, status labels, and the hidden-row indicator.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Hub: store the latest assistant result preview per session, backfill
existing rows, and expose lastAssistantMessagePreview in the session list.
- iOS: carry lastResult in the widget snapshot and add Running + Review
widget kinds. Running shows the task given to the agent; Review shows
unread assistant results. All home widget kinds reload together.
- Keep unread markers on running rows; Review excludes running so it only
shows finished results.
- Bump the local build to 29.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Carry the last-seen watermark into the App Group widget snapshot as
isUnread; the app refreshes the widget when read state changes.
- Mark push-updated sessions unread in the NSE snapshot path.
- Show a blue dot and blue timestamp for unread widget rows.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Hub: add a background APNs transport and a throttled widget-state
notifier; fan out silent content-available wakes to registered iOS
devices on widget-visible session changes.
- iOS: handle silent pushes and BGAppRefreshTask by refreshing the active
hub list and waiting for the App Group snapshot write; enable background
modes; register APNs tokens even when alert permission is denied.
- Store: repair a drifted fcm_devices schema missing push_key so
/api/devices/register no longer 500s (this also restores alert push
registration on the affected database).
- Document the silent-refresh fallback and its best-effort limits.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Include collapse overrides in the equatable list key; header taps were
reusing stale content until the next session update.
- Make the whole section header tappable and stop disabling it under
filters; filters now open groups by default but explicit taps work.
- Hide the redundant bucket subheader when In progress/Active has only
one bucket.
- Bump the local build to 18.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Add the HappyCoding app icon to the widget bundle and show it in the
home widget header (small/medium/large).
- Animate running/background rows with a right-aligned spinner in the
session list.
- Shorten the widget timeline fallback from 15 to five minutes; WidgetKit
remains a snapshot surface, not a live stream.
- Bump the local build to 16.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Add personal pins (My sessions), global/project pin separation, H5
running/pending and active/idle buckets, and collapsible project groups
with counts and selected-row expansion.
- Add local personal pin storage and list-specific display metrics.
- Add Settings > Appearance list text size (80-130%) and CLI/code font
size (9-17 pt), with Dynamic Type still applied.
- Keep search flat and filters force groups open; shared/archived/status
detail/preview-fold remain out of scope.
- Bump the local build to 14.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Group the large widget into running/active sections, show up to nine
sessions, and use the last human prompt as the subtitle.
- Show compact status icons plus web-style relative times, right-aligned.
- Mirror the per-install push key into the protected App Group container so
the Live Activity extension and NSE can decrypt while the data-protection
Keychain is unavailable on the Lock Screen.
- Redesign the Lock Screen card and Dynamic Island compact/expanded/minimal
states; Live Activities are now opt-in and recovered activities are ended
when disabled.
- Rename the app and widget display name to Happy Coding; bump the local
build to 10.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
Separate the content tree from the connection overlay so reconnection
banners cannot resize the transcript or trigger viewport layout handlers.
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
- add the WidgetSnapshot contract, App Group store, status/title builder
- write debounced snapshots from SessionListStore; NSE refreshes them
from decrypted pushes and reloads widget timelines
- add small/medium/large, accessory, and StandBy widget families with
attention-first rows and session deep links
- add App Group entitlements for the app, NSE, and widget extension
- document the snapshot data flow and privacy properties
- cover selection, title fallback, store round-trip, and push updates
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
Hub:
- add encrypted Live Activity token registry, registration API, and
update/end pushes over direct APNs
- enable only for HAPI_IOS_PUSH=apns; relay reports remote:false
- keep content-state encrypted with the per-install push key and a
domain-separated AAD
iOS:
- add ActivityKit coordinator, HapiWidgets extension, lock-screen and
Dynamic Island UI, encrypted content-state, deep links, and settings toggle
- cap concurrent activities and reconcile across hub switches
Docs/tests:
- native contract, privacy, READMEs, localization catalogs
- cross-language cipher fixture and hub/iOS unit tests
via [HAPI](https://hapi.run)
Co-Authored-By: HAPI <noreply@hapi.run>
Co-Authored-By: Claude Code <noreply@anthropic.com>
The Mac window's NSToolbar drops the custom .principal header, and the
hidden titlebar title also hid the navigation title. Use the system title
on Mac (as on the folded Duo) with a unified toolbar instead.
Build the app and notification extension for Mac Catalyst with the Mac
idiom, so Macs get a 1:1 native window instead of the scaled iPad app.
The Mac build keeps the run.hapi.app bundle ID and adds Mac entitlements
(sandbox, network, microphone, user-selected files, Mac APNs).
- Always use the split on Mac with a sidebar list style, denser rows,
14pt reading text, a hidden duplicate titlebar title, a minimum window
size and page-sized sheets; open web links in the default browser and
hide the unavailable live QR scanner.
- Add menu bar / hardware-keyboard commands: New Session, Settings,
Toggle Sidebar, a Session menu (settings, files, rename, export, stop,
previous/next session) and a guide link.
- Composer: Return sends on Mac (Shift/Option-Return for a newline, IME
untouched, first suggestion picked); dropped files and pasted images or
copied files become attachments on all platforms.
- CI builds Mac Catalyst and runs the navigation, split and transcript
geometry tests there; README documents the Mac build.
Profiling a 2,745-session hub while scrolling a chat showed the two
longest main-thread stalls (~80 ms) came from the session list, not the
transcript:
- The home list, covered by the pushed chat, still rebuilt and diffed
every row on each session change. A covered list now keeps its last
rows and refreshes once on return; a side-by-side split stays live.
- DiskCache encoded the 2 MB session snapshot on the main actor after
every burst of changes. Encoding now runs with the file write, off it.
After the change the trace has no session-list or snapshot work on the
main thread and no stall over 50 ms (max 38 ms, from Markdown cells
entering the screen).
The queued card's ScrollView is greedy, so one row still took the full
140 pt cap; measure the rows and size the card to them, capped at ~3.
Sends dispatched while the agent is idle no longer flash through the
queued bar while invokedAt is still null: the interactor records them
before the POST (the hub flips thinking on as soon as it queues), the
thread renders them, and after a 5 s grace a still-uninvoked row moves
to the bar as genuinely queued.
Any session change (even an offscreen one, e.g. a streaming agent)
re-ran every row's content (button, swipe actions, context menu), so one
update cost O(all rows): ~30 ms at 300 sessions in a Release simulator
build. Rows now compare only what they draw and skip unchanged content
(~4-6 ms, flat in the row count), and the minute tick redraws only each
row's age label instead of rebuilding the list.
Replace the full-width queued bar with a floating card stacked on the
composer, following the Codex/Claude Code/Cursor convention: one-line
rows with icon actions (bolt to steer, retry, cancel), tap a row to
edit, long-press for the full menu, and animated insert/remove.
- Collapse the empty, unfocused composer to one row; anything in flight
(focus, draft, chips, schedule, recording, scratchlist) expands it
without dropping focus.
- Bare 20 pt tool symbols with color-only states; only send/stop is a
filled 32 pt circle. Both scale with Dynamic Type.
- Fix the schedule menu reading as active: Menu tinted its relative
`.secondary` style.
- Stop uses the accent instead of alarm red.
- Liquid Glass card on iOS 26 (damped so text behind the short pill
stays out of the placeholder); hairline edge on earlier systems.
- ⌘↩ sends from a hardware keyboard, gated like the Send button.
- Transient SSE states appear only after 2s, as the iOS 26 nav subtitle
(a gray line with a spinner where no title is shown), so the list
never shifts on routine reconnects
- Only a failed list refresh is amber: "Offline · updated …" with Retry
Running sessions previously suppressed the ordinary unread dot, hiding
answers the user had not seen. Unread is now
manualUnread || contentRevision > readRevision on web, iOS and Android.
Web rows draw the running/background spinner beside the unread dot, and
unread outranks background status. Because session patches arrive before
the message they count, a chat following its visible tail hides its own
ordinary dot until the rendered acknowledgement lands (display-only; web
and iPad split view).
- Unread is a semibold title plus an accent timestamp, so the presence
badge stays the row's only dot and timestamps share one column
- Compact section header insets and default min header height
- README: the session list uses a large collapsing title
- Large collapsing title; rename the second section to Recent
- Swipe actions: pin/mark read (leading), archive or delete/reopen (trailing)
- Two-line agent icon with a presence badge (pulses while running) instead
of dimming disconnected rows and per-row spinners
- Larger timestamps, compact section headers, glass-friendly filter icon
- Drop the pixel-level row presentation tests
Track readable content revisions and durable read cursors in the hub. Acknowledge only rendered foreground tails, persist offline read operations, and preserve newer manual unread choices across devices.
Migrate existing history to a read baseline and cover pagination receipts with shared native conformance fixtures.
The hub's fork, rewind and export routes had no iOS UI. The chat's
menu now offers:
- Fork Conversation: when the agent advertises forkCurrent, the
session is live and no Codex turn is running (the web gate). The new
session opens in place of this chat.
- Rewind or Fork…: a sheet listing loaded user messages that have a
native point in metadata.conversationHistoryPoints. Each row offers a
confirmed Fork from Here or Rewind to Here, gated like web: forks
need an idle turn except on Codex, and rewinds always need one. The
actions live in a sheet, not on the transcript bubbles, because those
keep text selection and cached row heights. After a rewind the tail
resyncs with a history reset. An ambiguous native boundary (409
ambiguous_native_boundary_fork_safe) offers a fork from the same
point instead.
- Export Conversation…: saves the hub's lossless JSON payload through
the Files save panel, using the web download's filename. A large
export asks before retrying with force; a 413 explains the limit.
Markdown export stays web-only.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
iOS showed scheduled messages but could not create them. A clock
button now sits next to the composer's attachment button. It offers
the web presets (in 5 min, 30 min, 1 h, 4 h) or a specific time up to 7
days ahead. A chip above the input shows the pending time; its × sends
normally again.
Presets resolve when the user taps Send, so "in 5 minutes" counts from
the send. Validation follows the web: a 30 s grace period for past
times and a 7-day maximum.
A scheduled send is text-only and queued. The hub rejects scheduledAt
with attachments or steer, so the clock is disabled while attachments
are present and the steer menu is hidden. The schedule clears once the
send dispatches. Sending reuses the existing optimistic path, and a
retry keeps the row's scheduledAt.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Typing / in the iOS composer offered nothing; web and Android show the
session's commands. While the input is a lone /token, a list above the
card shows matching commands. A tap inserts "/name " so the user can
add arguments.
As on Android, the sources are metadata.slashCommands plus the
/slash-commands RPC. The RPC is fetched once, when the input first
enters slash mode, and its entries override bare names. A failed fetch
retries on the next /. There is no copy of the shared built-in table,
so the CLI stays the source of truth. Ranking follows web: exact,
prefix, contains, then near-miss typos. The list is hidden in
Scratchlist mode.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The iOS session list could only filter by machine. Web also has text
search, an unread-only filter and Mark All as Read.
- Search: a .searchable field ranks matches with the web relevance
rules, ported from sessionListSearch.ts via Android as
SessionListSearch. It searches title, worktree, summary, agent,
machine, path and id; every word must match, with */? wildcards.
Results keep the global/project pin sections, so the Pinned header
still applies.
- Filters menu:
- It now holds an Unread Only toggle that keeps the iPad's selected
chat listed.
- The machine section appears only with two or more machines, as
before; the menu itself shows whenever the hub has sessions.
- Mark All as Read advances the local last-seen watermarks after a
confirmation, as on web.
- Empty list: an empty search or filter result gets its own empty state
instead of "No sessions yet".
getSessionTitle moves into HapiKit so search and the rows share it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The settings sheet only had model pickers for Claude and Codex; other
agents could not change models after creation. Live sessions now load
/sessions/:id/<flavor>-models. Agy uses its machine catalog, since it
has no session route. As on web, grok and copilot skip
terminal-controlled sessions.
Rows follow web getModelOptionsForFlavor:
- Copilot, dsh, grok and kimi offer a Default or Auto row that clears
the model.
- Cursor's Auto is the explicit 'auto' wire id.
- Opencode, agy and pi only switch between listed models.
- An unlisted current model stays visible.
Pi resolves provider/modelId from metadata.piSelectedModel and switches
with a {provider, modelId} reference.
Efforts:
- Grok and OpenCode options come from their *-reasoning-effort-options
routes and reload after a model switch. OpenCode options that still
describe the previous model stay hidden.
- Pi shows the current model's thinking levels.
The sheet's load-failure flag now covers every catalog, not just Codex.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The create form only offered models for Claude and Codex, so every
other agent spawned on its default model. The form now loads the
catalogs web uses. Agy, cursor, dsh and pi read
/machines/:id/<flavor>-models. Copilot, grok, kimi and opencode probe
the same route with cwd once the directory exists; the view refetches
when the typed path turns out to exist.
Option rows follow the web builders:
- Copilot and cursor label the default row "Auto".
- Kimi and pi prefix the provider; pi values are provider/modelId.
- OpenCode marks the agent's own default.
- Cursor shows a flat list of ACP ids, without the variant drill-down.
Grok efforts and pi thinking levels follow the selected model, and an
OpenCode variant becomes the reasoning effort. Grok's auto permission
mode is hidden when the probe reports it unsupported. A restored
selection the catalog no longer offers falls back to Default, and
Create waits while a non-default selection is still being validated.
The spawn body sends model for all of these agents, and cursor always
sends it, with 'auto' as the default. Grok and pi send effort; opencode
sends modelReasoningEffort.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The API wrappers existed but iOS had no UI for these actions, while
web and Android did. The session row's context menu and the chat's …
menu now offer Rename for every session. Active sessions keep Archive.
Inactive sessions get Reopen and a confirmed Delete instead of
Archive, which the hub rejects with 409 for them.
The store applies renames and deletes optimistically. A failed rename
refetches server state; a failed delete restores the row and the
cached detail. Reopen marks the returned session active right away.
When reopen returns a superseding id, the chat seeds the new window,
moves the draft and navigates to it, as send-resume already does. A
reopen from the list also moves the draft and opens the session. A
422 shows the hub's message with the missing fields.
On iPhone, an authoritative removal (delete, or archive elsewhere) now
pops that chat, matching the iPad split, which already clears the
selection.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(codex): drop safe-yolo from Codex permission modes
Every Codex session runs on the native shared runtime (runCodex is
runSharedCodex), which rejects safe-yolo. The mode stayed in
CODEX_PERMISSION_MODES and six filters hid it again in the web picker,
composer, hub permission-mode route, hapi set-mode and both native
apps. API callers that validate against the flavor catalog (e.g. a
spawn with permissionMode: 'safe-yolo') still passed it through and
only failed inside the runner.
Remove it from the Codex catalog (shared, iOS, Android, fixture) and
delete the filters. autoReview is the native replacement. Safe Yolo
stays for Gemini, Kimi and Copilot.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(ios): Codex option list now has four modes
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
- Use the split navigation on foldable iPhones once UIHingeInteraction
(iOS 27.1+) reports a hinge: unfolded shows sidebar + chat, folded
collapses to the phone stack without recreating the chat.
- Side-by-side sidebar: drop the truncated inline "Sessions" title and
move the filter beside the hub menu (also fixes iPad).
- Folded outer screen: drop the home title so the empty bar disappears,
use the leading-aligned system title/subtitle in chat, minimize the
chat title on scroll, and let the composer sit 16pt closer to the home
indicator while the text field is unfocused.
- Qualify Markdown.Document, which is ambiguous with the iOS 27.1 SDK.
- test-transcript.sh picks the newest runtime that supports the device
type (iOS 27.1 only ships iPhone Duo); split tests run on iPhone Duo.
* feat(ios): add Russian locale
Add Russian translations to Localizable.xcstrings for all 591 catalog
entries (reusing matching Android translations and translating the rest);
zh-Hans and format placeholders are preserved.
* fix(ios): make Russian selectable and use correct plural forms
- add AppLanguage.russian ("ru"), its display name, the AppleLanguages
override, and register ru in the Xcode project's knownRegions so the
language picker can actually select Russian
- add Russian one/few/many/other plural variations for %lld sessions,
%lld attachments and %@, %lld sessions
- use count-agnostic phrasing for the other dynamic counts (tools, edits,
lines, chars, findings, notes, queued messages, questions, ...)
- correct Input/Output to Ввод/Вывод
- cover the new language case in ThemePrefsTests
* fix(ios): use Russian plural forms for the remaining count labels
* fix(ios): pass the step count to the plural localization
* fix(ios): pass tool counts to the plural localization
* fix(ios): keep the singular English tool-group labels
---------
Co-authored-by: mrudy <mrudy@local>