1.7.2 Changelog

New

Long sessions & compaction

• Agent-initiated compaction: the Session tool gains a compact action, so a compaction can be started at a tool boundary while a run is in progress and execution continues from there—no need to wait until the context crosses the line.
• Visible compaction: the progress ring spins and finishes smoothly while compacting; the compaction boundary card now includes token counts and a summary, making it clear how much context sits on either side of the boundary.
• Space reclamation for archived sessions: archived sessions support storage statistics and space reclamation, so long-accumulated session data can be released safely.
• Long-session stress testing: a new long-session compaction stress-test CLI (run through npm run, with scenario suites, JSON output, and assertion levels) for local regression of long-session scenarios.

Mini Tool platform

• Immediate follow-up API: the mini program session API gains ctx.sessions.promoteTurn, which moves a mini program's own queued message to the front of the queue along the same cooperative-interruption path as the UI's "follow up now"; it is paired with ctx.api.supports() detection so mini programs can stay compatible with older hosts.
• Session titles: mini program session information gains a title field, so mini programs can display the session name directly.
• Toolbar layout: the mini program toolbar scrolls horizontally and gains an icon-only mode, so narrow windows no longer squeeze the input area.
• Background subtask entry: mini program views get the same background subtask menu as the main session, letting them jump straight into background sessions they created, reusing the same unread and running-state badges.

Models & providers

• Custom request headers: model providers can be configured with custom request headers, making it easier to connect gateways or self-hosted services that need extra authentication headers.
• Model catalog update: updates the built-in model catalog and authoritative catalog, adding GPT-6.1 Sol; a session's model choice is now pinned, so reopening a past session restores the model it used at the time.

Diagnostics

• Main-process network console: a new network console means proxy and request problems no longer require digging through logs; database performance diagnostic logs are added alongside it.

Improvements

• Faster long-session opening: branches and entries are merged into a single pi round trip; the first screen reads only the tail window of a branch, and earlier rounds are fetched on demand.
• Leaner session file compression: on-disk compression of session logs now streams line by line instead of reading large files into memory all at once.
• Large code blocks no longer block typing: chat syntax highlighting now runs asynchronously in slices.
• Image format in preview: the full-screen preview title bar derives an image format chip from the MIME type, file extension, and file magic number, and shows nothing when the format cannot be identified.
• Consistent quota reporting: subscription quotas are shown consistently in Settings, the model menu, and the tray, with clearer reset times and plan display; an empty card is no longer shown when quota has not been fetched.
• Browse models while signed out: when signed out, all of Finch's available models are shown, narrowing to your own permissions after signing in.
• MCP presentation: refines the MCP service card and its capability list, aligns the height and empty state of the switch dialog, and collapses a fully empty connector list into a centered icon, a single line of guidance, and an add button.
• Mini program permission copy: permission declarations and incremental review copy are clearer, so a first-time grant is no longer mistaken for a "permission upgrade".
• Pin queued messages: queued messages can be pinned, with clearer action hints and queue-full messaging.
• Bash reclamation: Bash shells are reclaimed by run ownership, and a session-level force-kill entry point is added.
• Storage layering: configuration, rebuildable caches, and run state are split by purpose; legacy data such as model config versus model settings, session display state versus clearable caches, window size versus session restore state, automation tasks and their run records, log settings and mini program state, workspace context, and unread sessions and usage history all migrate automatically at startup, with a standalone CLI following suit. Files left over from the previous version are cleaned up according to the established list.

Fixes

Sessions, timeline & compaction

• Fixes tools being executed twice after reconnecting from a network hiccup: calls with identical parameters reuse the result that already succeeded before the reconnection, duplicated tool cards in the timeline collapse into one marked ×N, and side effects such as summoning multiple agents twice are eliminated.
• Fixes queue problems during compaction: follow-ups jumping the queue, queued messages arriving out of order, the queue being dispatched early while a compaction rerun is still in progress, and leaked state when compaction finishes.
• Fixes Unicode line separators (U+2028 / U+2029) causing dropped frames in session history IPC.
• Fixes a blank session timeline after a model timeout.
• Fixes link parsing in assistant replies, where bolded URLs and the markup immediately following them are no longer swallowed.
• Fixes image previews and relative links under Chinese paths: non-ASCII characters in Markdown links were double-encoded, so the main process could not find the file; relative links crossing between sessions are fixed as well.
• Fixes attachments that kept occupying the input box after a failed read, and draft attachments left behind after sending from Home.

Queue, follow-ups & input

• Fixes four issues in the follow-up path: bottom-stick scrolling briefly failing, the page bouncing back down when a reply completes, the pending-send button staying in a loading state after a follow-up is consumed, and submitting an edit accidentally sending queued messages.
• Adjusts the waiting policy for permission cards: in ask mode there is no longer a countdown that auto-declines, so approval can take as long as it needs; act mode keeps its two-minute countdown and timeout decline.

Models, accounts & mini programs

• Fixes custom models being misjudged as not supporting thinking, built-in providers falling out of sync with the remote catalog, local and remote providers being merged, messages failing to send after switching to a particular model, and OAuth sign-in from model settings.
• Fixes Finch Cloud account and subscription state falling out of sync, quota still showing after signing out, short-lived credentials being refreshed immediately, and quota differing between the tray and Settings.
• Fixes mini program subtasks showing unread after completion, archived sessions still showing unread, the panel fetching its list repeatedly, the full-screen page showing a duplicated hierarchy, a missing first breadcrumb level, and uninstalled bundled mini programs being reinstalled automatically.

Platform stability

• Fixes force-quitting a session restarting the pi runtime, the main process potentially launching twice in development, and the app failing to close after shutdown cleanup.
• Fixes incomplete file search results in large workspaces, the Workspace split still depending on old data files, the Space "more" state frequently failing, and other expanded spaces collapsing when one space is expanded.
• Fixes quotes and dollar signs in shell commands being expanded prematurely.

1.7.0 – 1.7.1 Changelog

New

MCP & mini programs

• MCP toolbox: the toolbox gains an MCP tab where you can add, edit, and manage your own MCP servers, and complete authorization when a service needs it.
• Clearer MCP status: service status badges, error messages, and authorization icons are unified, and the add-server form and list interactions are easier to follow.
• Mini program collaboration: mini programs can now observe the global waiting state and answer on the user's behalf, and they gain capabilities for reading the model catalog, switching models per turn, and specifying a model when creating a session.
• Shared collaboration foundation: mini programs get owner-isolated shared artifacts, documents, tasks, and handoff records, making it possible to build traceable multi-agent workflows.

Sessions & collaboration

• Subtask batch navigation: a subtask session's title bar can show the other tasks in its batch and switch between them from a dropdown.
• Side chat: the right panel can open an independent side chat that keeps the current workspace context without disturbing the main session's streaming output and interactions.
• Updated archive management: the archive page in Settings has been redesigned, and the automation panel can also show archived run sessions.
• Quick quote: the floating actions that appear on selected text are consolidated into a single, simpler quick-quote entry point.

Models, permissions & platform

• Model price multipliers: the model menu can display price multipliers so different models' costs are easier to compare; xAI brand icons and the login flow presentation are added.
• More flexible act mode: command gating decisions are unified; cleaning up declared temporary directories, project .temp scratch scripts, and compound commands now run more smoothly according to their actual risk, while high-risk operations are still called out explicitly.
• Configurable execution shell: the shell used for command execution can be configured on macOS, Linux, and Windows; on Windows that includes Git Bash, CMD, and PowerShell.
• Workspace rule chain: AGENTS.md rules are loaded automatically along the Git repository hierarchy, giving in-project collaboration more accurate context.

Improvements

• Startup and data migration: session, Space, usage, OAuth, extension, and workspace data migrations now all complete during startup, making upgrades from older versions more reliable.
• State persistence: workspace layout, sidebar preferences, editor drafts, settings, and session state are stored split by domain, so they restore more consistently across restarts and page switches.
• Permission explanations: permission cards highlight high-risk operations and use plainer wording; a wait timeout now states the model's reason clearly so it is not confused with a user decline.
• Privacy-friendly usage statistics: anonymous statistics are reduced to a startup signal and a local daily-active report, cutting unnecessary event collection.
• Interface details: refines hover states and padding on option cards, secondary timeline information, and the visual hierarchy of assistant bubbles and tool-call cards.
• Markdown link handling: links gain a context menu, and image syntax pointing at a non-image file now degrades to a file link that opens normally.

Fixes

Sessions, drafts & timeline

• Fixes Composer drafts and attachments being lost after a switch or restart, and draft directories lingering after attachments are deleted.
• Fixes being unable to keep sending after an edit-and-resend request loses its connection, and history messages that could not be edited or branched from after compression.
• Fixes the viewport being dragged back to a previous reading position after sending, editing lag in long sessions, and overlong titles that could not be viewed in full.
• Fixes a session view crash caused by incomplete tool-call summary fields, suggestion rows clearing the draft when sent directly, and missing subtask titles or completion summaries.
• Fixes annotations being submitted accidentally while confirming an IME candidate, accidental submission from the model settings input, and group dragging when selecting text during a sidebar rename.

Models, accounts & mini programs

• Fixes a session continuing to use the old provider after switching models, and professional mode not falling back to the model's recommendation when it lacks the requested thinking level.
• Fixes xAI login hanging while fetching the authorization code, the Finch Cloud account card loading indefinitely, and model counts and toggles displaying incorrectly when signed out.
• Fixes model usage being counted twice, price-multiplier badges not rendering when the multiplier is zero, and repeated tool calls by a model potentially forming a loop.
• Fixes the mini program panel being mispositioned when opened in a background session, truncated button labels, and incorrect MCP authorization icons for previously installed mini programs.

MCP, extensions & platform stability

• Fixes the same extension package being installed twice, old directories lingering after an installation is replaced, and activation before an MCP credential migration has finished.
• Fixes Grep failing on Windows first run because ripgrep was missing, and adds git worktree remove to dangerous-operation detection.
• Fixes the dev build losing its Dev suffix in the app name, errors when checking for updates on dev builds, and deep links and OAuth authorizations interfering across different build variants.
• Fixes the sidebar turning unexpectedly dark in the light theme, the model menu being mispositioned after a wait dialog, and the sidebar not following when the unread badge is clicked.

1.6.3 Changelog

New

Mini programs & extensions

• Appearance capability: mini programs gain ctx.appearance, which can set the theme, custom skins, font size, font family, and Home background once permissions.appearance is granted; the permission card and its trilingual copy are rounded out alongside it.

Models & providers

• Remote provider catalog snapshot: fetching a remote provider catalog snapshot is now supported, making model catalog state more reliable.

Improvements

• Model menu cleanup: groups models by source, unifies source-card corner radii, pinned-model hints, and empty-state layout, and adds diagnostics for where a model switch originated.
• Sidebar and theme details: the sidebar menu switches to a hamburger icon, timestamps and captions are tightened, and the light-theme active row and loading animations are refined.
• Emoji coverage: adds 142 commonly used local Twemoji assets covering animals, food, nature, sporting goods, and more.
• Log maintenance: expires old logs by date, caps individual file size, and shortens the detailed log retention window to seven days.

Fixes

Sessions & compression

• Fixes streamed content rolling back, trailing flicker, stale compression boundaries, and lost follow-up suggestions when switching sessions.
• Fixes compression boundaries disappearing after collapsing a round, stale operations clearing compression state, and queued messages appearing early during a compression rerun.
• Fixes send-then-navigate timing on Home and a stale open-identity cache when reopening the latest session after creating one in a container.
• Fixes incorrect attachment restoration on the first image-and-text message, duplicated action suggestions in an error state, and missing retry details in run status.
• Fixes a theme flash when creating a session and Space session list expansion state not being written to the database.

Models & Finch Plan

• Fixes Finch Plan model catalog refresh failures, accidental automatic switching during a refresh, delayed subscription status updates, and network fluctuations signing the user out.
• Fixes new sessions not restoring the last model, defaults not updating after a model switch, broken cross-provider switching when a model becomes unavailable, and invalid drop targets when dragging models.
• Fixes usage quota ordering, duplicated Finch Plan quota card titles, and log usage not updating continuously.
• Fixes an incorrect terminal state being overwritten after a cloud model interruption.

Mini programs, rendering & platform

• Fixes concurrent mini program update states overwriting each other and permission warnings displaying incorrectly.
• Fixes relative file links being rendered as blocked and local images failing to display on Windows.
• Fixes text-default emoji such as ✌️ not rendering correctly as Twemoji.

Appearance

• Corrects the custom skin --text-inverse derivation: instead of inverting RGB directly, it now blends white or black according to the theme base, preventing saturated dark colors from looking washed out or gray on accent backgrounds.