1DevTool Changelog

Weekly releases for 1DevTool — the AI code editor that runs Claude Code, Codex CLI, and Gemini CLI side-by-side. Every new feature, improvement, and fix, in reverse-chronological order.

Currently on v1.72.2 — Wait for a terminal to finish, then send the follow-up. Download the latest build →

v1.72.2

September 10, 2026

Wait for a terminal to finish, then send the follow-up

Ask for work that starts when another agent is done. With orchestration on, Agent Input now understands wait for terminal <id> to finish, then implement the findings — your follow-up is held until that turn really ends, and Mission Control shows every pending wait with a Cancel wait button. Alongside it: a macOS notch strip that answers "is anything waiting on me?" without raising the app, readable subagent history for the children Claude Code and Codex spawn, a folder-tree view over Git changes, Paste Into Terminals on any file in the explorer, and SQL Server connections over Windows Authentication.

The Conversation view's message rail on the right, one dot hovered so a rendered markdown preview of that prompt opens beside it with headings, numbered steps and inline code intact
The Git Changes dialog in folder-tree mode, 18 changed files nested under scripts, src/main/database/adapters and renderer/components with a per-folder staging checkbox on each directory row

New Features

  • +The notch tells you what your agents are doing (macOS) — a strip under the camera notch shows one brand mark per live AI terminal and one line of text with the most urgent state: 1 needs you in amber, else 2 working, else 2 idle.
  • +Hover the notch strip to expand a list — brand mark, terminal name, project, state — and click a row to jump to that terminal, switching project if it lives elsewhere.
  • +The marks carry state on their own: a working agent pulses, one waiting on you flares, an idle one dims. The strip only exists while an AI terminal is alive, and turns off in Settings → Notifications → Show agent status in the notch.
  • +Subagent history for Claude Code and Codex — a count beside the delegation badge opens a searchable list of the children your agent spawned; picking one shows its conversation, tool calls and results.
  • +Finished children stay in the list, so a delegation can be read after it ended instead of vanishing with the spinner. Follow log streams a running child, Earlier logs pages back, Latest jumps to the end.
  • +Copy ID and Copy follow-up for parent hand you a prompt to paste back into the parent conversation. Reading only — 1DevTool never starts, writes to, or steers a native child.
  • +Wait for a terminal to finish, then send the follow-up — with orchestration on, Agent Input understands wait for terminal <id> to finish, then implement the findings, and holds your prompt until that turn really ends.
  • +Chain several dependencies (wait for terminal <a> and terminal <b> to finish, then run the integration checks), or address one by its Codex or Claude session id when exactly one open terminal owns it.
  • +Mission Control lists the pending waits, what each is waiting on, and a Cancel wait button; Copy original prompt gets your text back if you cancel. Nothing is sent upstream — the waiting agent is never nudged.
  • +Waits survive pane remounts and project switches, and refuse the cases that cannot mean anything: waiting on yourself, duplicates, and cycles.
  • +Git changes as a folder tree — a toggle next to the filter box switches the Changes list between the flat list and a VS Code style folder tree, and the choice sticks the next time you open the dialog.
  • +A folder checkbox stages or unstages everything beneath it in one click, and shows a dash when only part of that folder is staged. Collapse all / expand all handles a change set spanning many directories.
  • +Paste Into Terminals — right-click any file or folder in the explorer and drop its path straight into a running terminal. The submenu lists every open terminal grouped by project, current project first, each row carrying its agent's brand mark.
  • +Picking a target switches project if the terminal lives in another one, expands a collapsed terminal pane, and focuses the terminal — so the path lands where you are looking.
  • +Preview a past prompt as rendered markdown — hover a dot in the Conversation message rail and the message opens beside it with headings, lists and inline code laid out, instead of one clipped line of raw text.
  • +SQL Server with Windows Authentication — pick the auth mode in the connection form and add a domain when you need one, instead of being forced onto a SQL login.
  • +Pasting a connection string that already says authentication=windows, trusted_connection or integrated security fills the form in correctly, and copying one back out keeps it.

Improvements

  • ~Quota Center shows the agents you actually use — cards for providers you turned off in Settings → Agents, or have never signed into, no longer take up room; Show all reveals them with a note saying why each was hidden.
  • ~Claude's per-model weekly caps get their own rows — the same "All models" plus per-model breakdown you see in claude.ai's plan-usage dialog, next to the plan-wide numbers.
  • ~The composer scrolls to the line you just made — ⇧⏎ continuing a list, and Tab / ⇧Tab indenting the last line, no longer leave the caret clipped behind the footer bar.
  • ~A collapsed project group tells you it holds the selected project — in that project's own colour, without expanding — and picking a project from the dropdown, a shortcut, or the dashboard scrolls its row into view.
  • ~Middle-click closes an editor tab, the way it does in browsers and VS Code.
  • ~The Classic and Horizon presets resize with no file open — the divider between the editor pane and your terminals is there from the start instead of appearing only after you open something.
  • ~A collapsed file explorer keeps its way back with nothing open — the toggle stays on the collapsed bar instead of squeezing the "select a file" hint into 32 pixels with no way to reopen.
  • ~Reader Mode's Stack layout keeps a multiline prompt as one section — an outline, brief or JSON paste no longer explodes into a card per line for Claude and every other native session, matching what Chat layout already did.
  • ~The Reformat button is gone from the terminal headers — it could redraw the live frame but never rebuild committed history at the pane width, so it promised a repair it could not do. Reader Mode and Conversation still reflow the same transcript properly.

Bug Fixes

  • -Fixed an AI answer appearing twice in the terminal, the two copies wrapped at different widths — a resize made the agent redraw its whole answer over history it could no longer erase, and pane changes that only re-measured the font were triggering it on ordinary tab switches. Resizes now only happen when the pane really moved.
  • -Fixed Ubuntu .deb installs failing to update — 1.71 offering 1.72.1 ended in "This release has no compatible download for your platform/architecture" and retried forever. The release feed now lists both Linux formats, so .deb installs update in place.
  • -Fixed a fresh Codex terminal stuck on "Waiting for session transcript…" — Conversation stayed empty after a prompt submitted from Agent Input even while Codex was visibly running tools; the first turn now binds without waiting for a late hook.
  • -Fixed a linked prompt sitting unsent in Codex and OpenCode — a delivered prompt could stop as [Pasted Content 8306 chars] or [Pasted ~8 lines] in the native composer, needing a manual Enter.

v1.72.1

September 7, 2026

Prompt history in the composer, and a cleanup that leaves your agents alone

The prompt you want to send again is now one click from where you are writing. A history button on the Agent Input toolbar opens Prompt History above the composer, and Use drops the past prompt in at the caret — replacing a selection, or landing blank-line separated beside what you have already written — so borrowing a prompt never costs you an unsent draft. Teams got their memory back too: a stopped writer can be recovered onto the same dedicated worktree it was already using, keeping dirty allowed changes where they are instead of starting a fresh checkout. Improve performance no longer stops running agents — it sleeps idle browser tabs and leaves every CLI and its screen history alone — and the AI tools are named on the bar again, with plan quota moved into the status bar as a strip of logos and percents for the agents you actually have open.

The Agent Input composer with a history button on its toolbar and the placeholder reading "Click here to view your Prompt History", so a past prompt can be reused without leaving the draft
The top bar with Orchestration, Sessions, Prompts, Memories, Skills, MCP and AI Diff shown as a centre group with their labels pinned instead of collapsed behind a sparkle menu

New Features

  • +Reuse a prompt you already sent without leaving Agent Input — a history button on the composer toolbar opens Prompt History above the draft, even when the composer is expanded or detached.
  • +Use drops the past prompt in at the caret, replacing a selection or sitting blank-line separated next to what you have already written — it never overwrites the unsent draft.
  • +Escape closes the dialog and leaves the composer standing, and Copy is still there when you opened history from the top bar instead.
  • +A stopped team writer can come back on the same dedicated worktree — closing a seat no longer forces its replacement into the project root or a brand-new checkout, so dirty, allowed changes stay where they were.
  • +From the team host, team capabilities reports what recovery can do and team recover brings the member back with its transcript handed to the replacement.
  • +Point a new terminal team member at an existing worktree and it is kept, not recreated; a missing path is refused rather than silently replaced.

Improvements

  • ~The AI tools stay named on the bar — Sessions, Prompts, Memories, Skills, MCP and AI Diff sit in a centre group with their labels pinned, instead of hiding behind a sparkle menu.
  • ~Remote moved up next to Activity, so pairing sits with the other live-state controls.
  • ~Terminals and plan quota moved to the status bar — quota is now a strip of logos and percents for the agents you actually have open, collapsing to the focused one when the bar runs out of room.
  • ~Activity is the button and its unread count; recent-event chips no longer take a strip of the bar.
  • ~Every AI entry point uses the same robot mark, including Agent Input, git's commit-message button, and Describe with AI.
  • ~Improve performance no longer stops your agents — it sleeps inactive browser tabs and leaves every CLI and its screen history running. Hibernate is still there when you ask for it, and only for an idle agent with a saved session.
  • ~A hibernated or failed terminal stays readable — Conversation keeps the session you were in, with Resume, Retry and Close in a footer instead of a live composer on top of a blank pane. Resume is a button that runs once; Close cannot start it by accident.
  • ~Reduce Motion actually stills the empty-state team glyph — with macOS Reduce Motion on, the travelling spark no longer loops in every pane that has no terminals.

Bug Fixes

  • -Fixed Hibernate and Improve performance stopping work that was still running — an agent that looked quiet could be mid-tool; generic relief no longer kills CLIs, a stop that cannot prove the process is idle is refused, and Resume no longer races a second kill against the session it is restoring.
  • -Fixed Windows terminals requiring Ctrl+Shift+C to copy — with text selected, Ctrl+C copies it; with nothing selected, the same key still interrupts PowerShell. Ctrl+Shift+C still copies.
  • -Fixed a machine terminal typing its startup command before the shell was ready — the first prompt used to eat the line as keys and execute a fragment of it, including an SSH key path. Startup commands now wait for the shell's first prompt.
  • -Fixed a Mac machine listing empty folders — listing files through the login shell aborted on unmatched globs, so the file tree looked empty even when the host had content.
  • -Fixed tmux on a Mac machine reporting env: tmux: No such file or directory — a Native Terminal that needs tmux now sees the same PATH as your login shell, including Homebrew.
  • -Fixed the empty-state team glyph keeping a core busy under Reduce Motion — the override shipped in 1.71 and 1.72 never won, so the animation ran in most panes.

v1.72.0

September 7, 2026

Every Machine, Every Pipeline

Every machine your projects live on, and every pipeline they build through, now have a home in the workspace. Machines is one roster for this computer, a VPS, a Mac mini, a VM or a paired 1DevTool install — each row reporting whether it is reachable, what OS and shell it runs, and which AI CLIs are installed on it. Add Terminal's This agent runs on control decides where the CLI process actually runs, locally against a mount or over SSH on the machine itself, per project and overridable per terminal. A new Repo panel puts GitHub and GitLab pipelines and issues beside your terminals, with re-run, cancel, searchable job logs, and a Send to AI that hands an issue and the comments you pick straight to an agent's composer. Kiro CLI and Oh My Pi join as built-in agents, and a new Extensions tab lets you switch off the panels you never use — restoring your exact layout when you switch them back on.

The Add Terminal dialog with the "This agent runs on" control set to This computer instead of kai-m1.local, explaining that the CLI runs locally with your own login and MCP servers while reading the machine through a mount, above the AI Agents grid
Project rows in the sidebar carrying kai-m1.local machine badges, with the status bar pill reading Desktop · kai-m1.local so the machine a project lives on is visible without opening it

New Features

  • +Add a machine once, and point projects at it — a VPS, a Mac mini, a VM, or a paired 1DevTool install lives in Settings → Machines alongside This Computer, which is always there and never has to be created.
  • +Every machine row says what is actually on it — whether it is reachable, how far away it is, the OS and shell, which AI CLIs are installed and at what versions, plus git, ripgrep, tmux and Node.
  • +Machines refuse to be deleted out from under a project — rename, remove, pick an SSH key, reconnect or mount one, and a machine still in use says so instead of orphaning the project.
  • +Add an SSH host from a picker fed by your own ~/.ssh/config, so aliases, ProxyJump and IdentityFile carry over instead of being retyped.
  • +Choose where your agent actually runs — Add Terminal's This agent runs on control offers This Computer (the CLI runs locally with your own login, quota, memory, skills and MCP servers, reading the machine's files through a mount) or On the machine (the CLI runs over SSH on the machine itself). The app never switches between them on its own.
  • +Agent placement is per project, overridable per terminal, and an agent that is not installed on the target machine is greyed out with a reason instead of failing at launch.
  • +Every surface says where you are — a machine badge on the project chip in the top bar, project rows grouped by machine in its dropdown, a chip in the sidebar, and a machine pill in the status bar with reconnect.
  • +The file tree tells you when a machine is unreachable — offline and sign-in-required states with a Reconnect button, instead of an empty folder.
  • +Add Project opens on the machine you are looking at, with its folders already listed and a new Machine picker to switch back to This Computer or any other saved machine.
  • +Paired computers and SSH hosts are one list now — the Devices tab folded into Machines, and Machine Terminal in the Add Terminal menu is now called Native Terminal.
  • +A new Repo panel with Pipelines and Issues sub-tabs, for GitHub and GitLab, resolved from the git remote of the project you have open.
  • +Watch CI without leaving the workspace — runs for the current repo and branch, expandable jobs, and the checks on HEAD, refreshed while you are looking and paused when you are not.
  • +Re-run, re-run failed jobs, and cancel a pipeline from the panel, each confirmed inline before it fires.
  • +Open a job's log in a searchable viewer with its colors intact, copy it, or jump to it in the browser.
  • +Read and write issues — filter by state, open one to read its rendered Markdown and comments, then create, comment, close, reopen, and change labels and assignees.
  • +Send an issue or a comment straight to an AI agent — the text lands in an agent's composer, with the issue URL and the comments you tick included as context, instead of your clipboard.
  • +Opt in to a background CI watch per project — a run that goes from passing to failing raises a desktop notification that focuses the project when clicked. Nothing is watched by default.
  • +Kiro CLI and Oh My Pi are built-in agents — AWS's Kiro CLI and Oh My Pi appear in Add Terminal, the empty-state cards, Change AI, Resume, Prompt History, Memory, MCP and skills, with their own logos.
  • +Those sessions resume after a restart, each through its own resume command, and titles you set are written back to the CLI's own store.
  • +A new Settings → Extensions tab switches off Browser panel, HTTP client, Database client, Mobile emulator, ToolBox, AI Diff and Repo — removing their panels from every layout and stopping their tools reaching your agents and your phone.
  • +Your layout and your data survive the round trip — saved layouts are never rewritten for a disabled extension, so turning one back on restores every panel where it was, with its connections, saved requests and browser tabs intact.
  • +Right-click any right-sidebar tab strip for the same extension switches without opening Settings.
  • +New installs start lean — just the browser — instead of seven panels you did not ask for. Existing installs keep every extension on.
  • +The SQL Server driver downloads on first use (3.5 MB) and is checked against a signature built into the app before anything is installed, so everyone who does not use it stops paying for it in the download.
  • +The right panel with no project open explains itself — the project list, Open folder…, and a link to Extensions, instead of an empty column.
  • +The toolbar and the header merged into a single 36px bar — 26 evenly-weighted slots became 10 buttons, with Orchestration, Session, Tasks, Terminals and Commands out front, AI surfaces behind a sparkle, and workspace and deploy tools behind ⋯.
  • +Labels collapse when the bar actually runs out of room, not at a fixed window width, and your project name stays whole at every size.
  • +A calmer project sidebar — root groups read as filled sticky bands, nested ones as small eyebrows, the group color runs down its subtree as a thin spine, and project names line up whether or not a project has an avatar.
  • +Five quick actions live in the project row itself, on hover instead of a permanent second row, and the branch label only appears when it is worth reading — a project on main does not spend a line saying so.
  • +Settings tabs are grouped into Workspace, Agents & connections, Development and Application, the AI tab is now Agents and owns the AI-terminal settings that used to be buried in Terminal, and every existing deep link still lands on the right pane.

Improvements

  • ~The five AI surfaces finally look like one product — Resume, Prompt History, Memory, Skills and MCP share one typeface, one filter bar, one control height and one row shape.
  • ~Session rows dropped from about 80px to 56px, so eight sessions fit where five used to, with row actions fading in on hover or focus instead of sitting there permanently.
  • ~Name a terminal, or copy what identifies it — right-click a pane for Set Title…, Copy Session ID, Copy Terminal ID and Copy Pane ID, and renaming writes the same title back to the agent's own session.
  • ~Reading Mode's column follows your font size instead of a fixed pixel width, with a Width stepper in the toolbar to nudge it.
  • ~The terminal launch screen was rebuilt — agent cards are single rows that fit more per column, three stacked sections collapsed into one footer bar, and a machine with no agent CLI installed says so instead of offering a card that cannot launch.
  • ~Zooming the terminal font is half the work it was — every visible pane used to be resized twice per step, making every full-screen agent redraw itself twice.
  • ~Process names on Windows no longer depend on wmic, which Windows 11 24H2 removed — the Ports panel and process attribution read from a native snapshot, and a failed read says so rather than reporting an empty machine.
  • ~A flooding pane you are actually watching stays readable — output from a storming terminal reaches the screen in about 25ms instead of 90ms when that pane is visible, while off-screen panes keep the batched schedule.
  • ~Terminals can recover their GPU rendering — if the window falls back to software rendering, the Resource Manager names the reason and the app tries once to get GPU rendering back when you return.
  • ~Switching to the project you are already on writes nothing to disk, project avatars use a fraction of the memory they did, and renaming a project no longer stalls on every keystroke.
  • ~A hidden terminal pane stops blinking its cursor deterministically, rather than relying on the browser to notice.
  • ~The "1DevTool paused" notice is a hint now, not a habit — it shows once, then progressively less often (10 minutes, then 20, 40, up to an hour), never for a stall that happened while the window was hidden, never takes focus, and has a Don't show again link.

Bug Fixes

  • -Fixed Conversation showing you walls of text you never typed — background-task completions, slash commands and interrupts that Claude Code records as user turns rendered as screen-tall "You" bubbles; they fold into one-line rows you can expand.
  • -Fixed Reading Mode wrapping earlier when you made the text bigger — the document column was pinned to a fixed pixel width, so a larger font meant fewer words per line.
  • -Fixed full-screen agents redrawing twice on every terminal font change — a single zoom step sent two size changes to every open pane.
  • -Fixed the Resource Manager popover opening on its own after using Open Resource Manager on a stall notice once.

v1.71.0

September 5, 2026

A release spent making the app stop eating your machine

A release spent making 1DevTool stop eating your machine. The Resource Manager now splits CPU and memory into three honest buckets — 1DevTool, the terminals you started, and other apps — that add up to what your OS reports, with an attention strip naming the pane that is actually the problem and Copy diagnostic report for a bug report. When the window stalls, a red dot in the status bar says so and a toast afterwards explains it: "1DevTool paused 6.2 s while system CPU was 98% — top: claude, node", with Recent stalls this session under Settings → General → Diagnostics. Underneath: switching layouts no longer rebuilds every terminal, a flood of output in one pane no longer freezes typing in another, idle agents hibernate, other projects' browser tabs unload, and a macOS install drops from 709 MB to 360 MB.

The Resource Manager panel splitting the machine into three buckets — 1DevTool at 24.9% CPU and 1.51 GB, Terminals · 1 at 1.3% and 1003 MB, and Other apps at 104.8% and 20.54 GB — above the Machine readings, Busiest other apps, and a Recent stalls entry reading "Main paused 3.1 s in command git while system CPU was 25%"

New Features

  • +See what is actually using your machine — the Resource Manager splits memory and CPU into 1DevTool, the terminals you started, and other apps, and the three add up to what your OS reports.
  • +The panel names the split your OS hides — Task Manager nests shells, node and claude under the 1DevTool row and Activity Monitor spreads the app across “Helper (Renderer)” rows; each bucket carries a one-line explanation of what it really contains.
  • +Expand a bucket to see what is behind the number — the individual terminals and processes, each with its own CPU and RSS, named the way its tab is.
  • +An attention strip for the terminal that is the problem — a pane pinning the CPU or flooding output gets a row naming the reason (“62% CPU”, “8.4 MB/s output · throttled”) with Reveal to jump straight to it.
  • +Act on it without leaving the panel — behind the row’s ⋯: Pause output, Lower priority, Hibernate, or Close.
  • +Copy diagnostic report — one click puts a Markdown summary of the buckets, top consumers, GPU rendering status and recent stalls on your clipboard, ready to paste into a bug report.
  • +The machine, in plain numbers — RAM used, Available, Swap, CPU load against your real core count, WebGL rendering status, and a Busiest other apps list so an expensive browser is named rather than blamed on us.
  • +When the app freezes, it tells you why — if the window stops responding a red dot appears in the status bar, instead of leaving you guessing whether the app died.
  • +A short toast once it comes back — “1DevTool paused 6.2 s while system CPU was 98% — top: claude, node”, so a freeze is attributable instead of mysterious.
  • +Recent stalls this session — listed in the Resource Manager and under Settings → General → Diagnostics, alongside the stall threshold and an opt-in for sending stall reports.

Improvements

  • ~Switching layouts no longer rebuilds every terminal — grid, columns, rows, single, vertical tabs, chat and vibe keep the panes you already had open instead of flashing Opening terminal on each one.
  • ~Jumping back to a recent terminal is instant — a pane you just left is still there, with its scroll position and selection, instead of replaying from scratch.
  • ~More terminals stay sharp at once — GPU rendering covers a larger set of panes, and a pane that is off screen gives its slot back so a busy workspace does not fall over to a blurry fallback.
  • ~A flood of output in one pane no longer freezes typing in another — that terminal’s output is batched and the pane says so, while every other pane stays at full speed. Nothing is dropped: the full output is there when the flood ends.
  • ~A workspace with dozens of terminals stays responsive — idle CPU and memory stay flat instead of climbing as you keep adding agents.
  • ~Terminal history is bounded by memory, not just line count — a very wide pane keeps fewer rows for the same memory, and widening a pane never throws away history already on screen.
  • ~Other projects’ browser tabs unload while you are away — after you leave a project its browser tabs release their memory once idle (Settings → Browser, 15 minutes by default). Tabs with unsaved fields, playing audio, driven by an agent or pinned Keep awake stay loaded, and returning wakes the active tab.
  • ~Idle agents can hibernate — an agent sitting at its prompt for half an hour shuts its CLI down and keeps the tab, showing “Hibernated — press any key or click to resume”. Never mid-turn, never a seat on a running team. Right-click for Hibernate or Hibernate all idle.
  • ~Huge HTTP responses stay out of your project file — a 20 MB response body no longer bloats the saved workspace; it lives in a bounded on-disk cache and loads on demand when you open the JSON view or copy it.
  • ~When the machine runs out of memory, 1DevTool gets out of the way — background polling pauses, idle browser tabs sleep sooner, and terminal processes drop to a lower CPU priority so typing stays responsive. Switch it off under Settings → General → Diagnostics.
  • ~The app is less than half the size — a macOS install drops from 709 MB to 360 MB, with roughly a third of the files it used to ship, so downloads and updates finish quicker.
  • ~Browser automation uses the Chrome or Edge you already have, with a clear message telling you what to install if neither is there.
  • ~Searching prompt history and opening the app no longer block the window — database work and startup checks moved off the thread that draws the UI.
  • ~Mission Control no longer steals the composer — when it auto-reveals beside Agent Input it borrows only the width the map needs, and gives the rest back the moment you switch panels or drag the splitter.
  • ~Diagrams match the rest of the workspace — Mermaid charts in the markdown preview and graph editor use a quieter palette, and loud author colours are toned down so nodes stay readable in both themes.

Bug Fixes

  • -Fixed the microphone recording silence in every terminal — Claude Code’s /voice, sox, ffmpeg and anything else that records worked in other terminal apps and captured nothing here, with no permission prompt and 1DevTool absent from the system’s Microphone list. The app now requests microphone and camera access properly, and Settings → General shows the status of both with a link to the right OS panel.
  • -Fixed “Description with AI” saving models the CLIs reject — a drafted team could come back with seats set to model names no CLI accepts, so starting it failed on the first turn with “model not found”. Model names are now checked against what each agent really offers, and a seat is left on its default rather than given a name that cannot run.
  • -Fixed Codex sitting on a Pipeline role card instead of joining the pipeline — activating a live-terminal Pipeline reported success and showed the links, but Codex kept the role card unsent in its composer, so later hand-offs found a busy composer and the agent looked like it “doesn’t link”. Role cards are now confirmed as accepted before the link counts as delivered.
  • -Fixed Conversation hiding Claude’s multiple-choice question after a real turn — once the agent had printed a build, a test run or any long stretch of output, the “agent needs your answer” card vanished even though the Terminal tab still showed the picker. The card now appears for those questions, including options that carry a two-column preview.

v1.70.0

September 2, 2026

Custom agents, and terminals you can select

Any coding CLI can now be a first-class agent in 1DevTool. Register a fork or a brand-new CLI in Settings → Terminal → Commands, give it a name, colour and logo, tell 1DevTool where it stores its sessions, and it appears in Add Terminal, Change AI, Resume, Prompt History and Memory exactly like Claude or Codex — no release required. Terminals also gained a selection mode: right-click Select, tick the panes you care about, then merge their transcripts into one agent or close them together. The status bar grew a worktree picker, and Reading Mode finally puts markdown and the terminal in a real resizable split instead of a dock that covered the page.

The Customize Startup Command dialog registering the omp CLI as a custom agent, with Run mode set to AI agent and the Agent integration panel showing accent colour #7C6AE0 and a Full-screen TUI screen mode
Edit Startup Command showing a custom agent set to Inline transcript with Pi-style sessions auto-detected at ~/.omp/agent/sessions and two recent sessions previewed before saving

New Features

  • +Custom agents — plug any coding CLI into the workspace — register a fork, a brand-new coding CLI, or a command you already launch by hand as a first-class agent, without waiting for a 1DevTool release. Open Settings → Terminal → Commands, mark the command as AI agent, and expand Agent integration.
  • +Give it a name, an accent colour and a logo — the logo is downscaled to 64px and stored with your settings; without one, the first letter sits on the accent colour (the default is #7C6AE0).
  • +A custom agent shows up everywhere a built-in one does — Add Terminal, the empty-state launcher cards, Change AI, Prompt History, Resume and Memory all render it with its own logo and colour.
  • +Sessions survive a restart — point 1DevTool at where the CLI writes its transcripts (Pi-style sessions, Claude-style projects, or a flat JSONL folder), then press Test to see real sessions listed before you save. Open reveals the folder in your file manager.
  • +Resume, continue and the agent's own picker are templates you fill in once — {cmd} --session {id} to resume by id, {cmd} -c to continue the latest, and either a slash command typed into the running TUI or a CLI flag such as {cmd} -r for the agent's own picker.
  • +Pick how the terminal treats the screen — Full-screen TUI lets the agent own the alternate screen and its own scrolling (OpenTUI-style, the safe default); Inline transcript blocks the alternate screen so scrollback stays in 1DevTool, the way Claude and Codex behave.
  • +Name the memory files yourself — list the per-project files (for example AGENTS.md) and an optional global file that the Memory panel should read and edit for that agent.
  • +Custom agents sit in the same Settings list as the built-ins — same Show toggle, same drag order. Built-in names are reserved, so you cannot register over Claude, Codex or any CLI 1DevTool already ships. Deleting the command asks whether to drop the registration; terminals already open keep working.
  • +Select terminals, then merge or close them — right-click a terminal and choose Select. Checkboxes appear on every visible tab and pane, including mosaic tiles, and the one you clicked starts checked. Checking a box never switches the active terminal.
  • +Merge several sessions into one agent — the action bar combines the selected transcripts into a single prompt and hands it to an existing AI terminal or a new one. The dialog lists each source with its message count, and warns when a terminal could not be attributed to a session so you know exactly what will and will not contribute context. Optionally close the merged terminals afterwards.
  • +Close a whole selection in one step — or right-click a project and choose Close All Terminals, with a confirm step that still leaves Reopen Closed Terminal intact. Select all, Done and Escape leave selection mode.
  • +Switch worktrees from the status bar — when a project has linked worktrees, the status bar names the active checkout with its colour and primary/locked/dirty marks, and switches without opening Git Client. Switching brings forward a terminal that already belongs to that worktree instead of leaving you on a tab from another checkout; a worktree with no visible terminals shows its own empty state.
  • +Markdown and the terminal are real side-by-side panes — Reading Mode no longer covers the document with a floating dock. Pick 5 : 5, 3 : 7, 7 : 3, markdown above the terminal, or terminal above markdown; drag the sash to resize, and the layout you pick is remembered.
  • +A diagram lightbox you can actually use — resize the source pane, collapse it to a rail, and double-click the sash to reset.

Improvements

  • ~Switching layouts no longer rebuilds every terminal — grid, columns, rows, single, vertical tabs, chat and vibe keep the panes you already had open instead of flashing Opening terminal on each one.
  • ~Jumping back to a recent terminal is instant — a pane you just left is still there, with its scroll position and selection, instead of replaying from scratch.
  • ~More terminals stay sharp at once — GPU rendering covers a larger set of panes, and an off-screen pane gives its slot back so a busy workspace does not fall over to a blurry fallback.
  • ~A workspace with dozens of terminals stays responsive — idle CPU and memory stay flat instead of climbing as you keep adding agents.
  • ~Mission Control no longer steals the composer — when it auto-reveals beside Agent Input it borrows only the width the map needs, and gives the rest back the moment you switch panels or drag the splitter.
  • ~Diagrams match the rest of the workspace — Mermaid charts in the markdown preview and graph editor use a quieter palette, and loud author colours are toned down so nodes stay readable in both themes.

Bug Fixes

  • -Fixed Conversation hiding Claude's multiple-choice question after a real turn — once the agent had printed a build, a test run, or any long stretch of output, the agent needs your answer card vanished even though the Terminal tab still showed the picker. The card now appears for those questions, including options that carry a two-column preview.

v1.69.0

August 28, 2026

AI Diagnosis and the Router, rebuilt

When an agent gets stuck on a quota wall or a sign-in prompt, another AI now reads its screen and tells you what's wrong — and hops the seat before the long timeout ever runs out. After about two minutes of a terminal failing to accept work, an installed checker AI reads that terminal's screen text and classifies the problem; a stuck team seat that has backup agents fails over on its own instead of waiting out the ten-minute timeout, and every other terminal gets the diagnosis as a notification. This release also rebuilds the Router as a plain match table — one row per kind of work with its own agent and model picker, read By work or By agent — lets you save an ad-hoc @agent team as a reusable chart, adds a play-to-learn Set-up screen that teaches the composer and the Router by running them, and gives AI file changes a Preview Files or Review mode.

The Agent Orchestration AI Diagnosis settings with 'Diagnose & restore stuck agents' toggled on, a 'With AI (recommended)' restore mode, and an ordered checker-agent chain of Claude then Codex under Advanced → AI Diagnosis
The rebuilt Router Defaults as a match table — one row per kind of work (test, review, browser, docs, research, debug, explain, advise) each with its own agent and model picker, and an 'advise → Claude / claude-sonnet-5' rule open in the side panel

New Features

  • +AI Diagnosis — find out why an agent is stuck — when a delivery to an agent terminal keeps failing for about two minutes, an installed checker AI reads that terminal's screen text and classifies the problem: a quota wall, an upsell modal, a sign-in prompt. A chart seat that has backup agents fails over on its own instead of waiting out the full ten-minute timeout; every other terminal gets the diagnosis as a notification so you can act. A healthy agent is never checked, a busy one mid-task is never mistaken for stuck, and reading a stuck screen takes a one-time consent because a screen can hold secrets. The toggle, the consent and the checker chain live under Advanced → AI Diagnosis.
  • +You choose which agents do the checking — an ordered checker chain in the dashboard: an agent that isn't installed, hits its own limit, or times out simply hands off to the next, and the stuck terminal's own provider is always skipped. Checkers run headless with restricted permissions in a neutral folder — they classify the screen and never touch your project.
  • +The Router, rebuilt as a match table — Defaults is now one plain row per kind of work (plan, test, review, browser, docs…), each with an agent picker and, once you pick one, a model picker right beside it. See it two ways: By work lists every kind of work and who takes it, By agent groups the rules under the agent that owns them; anything you haven't assigned sits in an Auto pool. Ask for a kind of work without naming anyone and the host routes it for you — and a guided Tutorial now plays the Router end to end.
  • +Save a prompt as a reusable chart — when you fan a prompt out to several @agents, the mode banner can now Save as chart, promoting the team you just described into a reusable Builder chart you can reopen anytime. What you saw in the send preview is exactly what gets charted.
  • +Learn the composer and the Router by watching them — a Set-up screen in the orchestration dashboard teaches by playing the one flow every new user needs: type a task, @-mention the terminals that should help, press Enter, and watch them start as a team while Mission Control lights up. Reduced motion gets a stepped walkthrough with every stage still shown.
  • +Choose how AI file changes appear — a new AI diff mode. Preview Files treats every change as accepted and just lists which files changed with added and removed line counts; Review keeps the per-file Accept / Reject controls with Keep / Undo overlays right in the editor. Switch between them anytime in Settings → IDE.

Improvements

  • ~Show or hide each project's terminals in the sidebar — a per-project terminal toggle that remembers your choice across relaunches and stays in sync across every sidebar; the header rearranges itself as you resize the sidebar so the control never gets cramped.
  • ~A seat that fails over stops holding onto its queued work — when a seat hands off to a backup, everything still waiting to reach the dead terminal now fails right away with a clear reason, instead of sitting in the queue until the timeout runs out.

Bug Fixes

  • -Fixed double-clicking a saved team opening every seat twice — a four-role chart could come up as eight terminals, and more clicks stacked more. Opening a team is now guarded against the double-click, the card is disabled while it works, and it shows how many terminals have actually appeared.
  • -Fixed a crashed agent seat that never handed off — an agent that dies on its very first turn (a seat that hit a provider billing wall, for one) now fails over to its backup within seconds instead of leaving the whole team stalled.
  • -Fixed a prompt swallowed while an agent was still signing in — the app now waits until an agent is genuinely ready before it counts a send as delivered, so a prompt sent mid sign-in isn't lost.
  • -Fixed every terminal failing to open after a startup hiccup — a single failure while the terminal engine was starting could brick all terminals at once; startup now recovers instead of hanging.
  • -Fixed broken screenshots in Software Update's release notes — release-note images now retry when a load fails and, if they still can't load, turn into a link you can click to open in your browser instead of leaving a broken-image icon.

v1.68.0

August 27, 2026

Backup agents and Task Improvement

A dead agent no longer takes the run down with it. Give any seat on a Charts org chart or pipeline an ordered list of backup agents — if the primary process crashes, runs out of quota, or an API gives up, the next backup opens in that seat and inherits the conversation so far, so the work continues instead of stopping. This release also lets you preview a whole agent team as an animated walkthrough before a single terminal opens, sets up your installed agent CLIs on first run behind one Yolo-or-Manual permission choice, and turns Tasks into something you can edit — a Markdown editor with Write Preview Split and a Rewrite with AI bar that drafts changes you approve before they are saved.

The Charts builder's Assignment panel showing a seat's ordered Backup agents list — 1 Claude on the Claude Opus-5 preset, 2 Cursor on the cursor full preset — that take over automatically if the seat's terminal dies
A task detail dialog with a Rewrite with AI bar reading 'Change the task name to Demo Orchestration', a Grok CLI agent picker and a Rewrite button that drafts the change for you to approve

New Features

  • +Backup agents — a seat that fills itself — give any seat on a Charts org chart or pipeline an ordered list of up to three backups. When the primary agent's process actually dies — a crash, an out-of-quota exit, an API that gave up — the next backup opens in that seat and receives the conversation so far, so it picks up where the last one stopped. Each backup can carry its own Terminal when starting preset, and closing a tab or quitting yourself never counts as a failure.
  • +Switch to a backup by hand — when an agent is visibly broken but its process is still alive, one click on the live seat hands it to the next backup and carries the transcript over. Active teams read backup 1 · took over from codex, so a result from the backup never reads as one from your first choice.
  • +Set up your agents on first run — a Set up your agents card on the welcome screen finds the agent CLIs installed on your machine and lists each with the exact command it will launch, behind one permission choice: Yolo launches each agent with its full-permission flag, Manual launches the plain CLI so it asks before editing a file. Only the commands 1DevTool generates ever change — your saved presets and typed commands are left alone.
  • +Preview a team before it opens — every orchestration card in Add Terminal now has a Preview that plays the chart step by step: a spark rides the wire on each beat, seats change state (briefed, working, reporting, done), and a caption says what just happened — so you see how four terminals will talk to each other before they appear. A preview only explains; Start is still one deliberate click.
  • +Edit a task, or have an AI rewrite it — open a task's detail dialog to edit its title, body and acceptance criteria, with the body in Markdown across Write / Preview / Split. A Rewrite with AI bar takes an instruction like tighten this or add acceptance criteria; the reply lands in the editor as a draft you read and Save yourself, and a Rewrite with picker chooses which installed AI CLI does the rewriting.

Improvements

  • ~The composer stopped asking you to speak orchestration — Agent Input's three setup switches (Hierarchy / Pipeline / Defaults) are replaced by a small drawing of the thing itself: one brief fans out, every agent works at once, one answer merges back. Everything those switches did already lives in Charts, the dashboard and the mode banner above the composer.
  • ~The launch screen fits the pane it's in, not the window — a terminal pane that's short because the editor is open above it, or narrow because another terminal sits beside it, now picks its own column count and scrolls the agent cards inside itself, so Agent teams, the plain shell and the layout picker stay on screen no matter how many presets you've saved.
  • ~Each agent card shows the command it will run — the launch screen's cards carry the actual startup command instead of a maker tag, which moved to the tooltip.
  • ~Click a terminal in Resource Manager to open it — a row in the per-process breakdown now switches project, unhides a collapsed pane and brings that terminal forward.
  • ~Orchestration sits next to the agents it opens — Add Terminal's tabs now read AI Agents · Shells · Orchestration · Machine Terminal.

Bug Fixes

  • -Fixed Enter and Send doing nothing with an @agent chip in the composer and orchestration left on Auto — a send that can't be routed now just sends normally, and a genuinely blocked send says why, right next to the Send button.
  • -Fixed a prompt sent from your phone waiting for a second Enter — Remote now proves the whole prompt arrived and that the agent took it before it confirms the send, instead of confirming as soon as the keystrokes left the pipe.
  • -Fixed the phone composer staying locked after a slow or dropped send — a lost connection settles straight away, your draft is kept, and the box is usable again without reloading the page.
  • -Fixed a reconnecting phone risking a double send — a retry after a dropped connection now replays the same request rather than sending the prompt a second time, and a send that may have half-landed fails fast with a recovery instruction instead of hanging.
  • -Fixed @file and $skill prompts from Remote being eaten by the agent's own picker — a prompt carrying one now arrives as pasted text, so the agent's autocomplete doesn't swallow the Enter that submits it.

v1.67.0

August 26, 2026

Draw your agent team

Every chart your agents run now lives in one library. Charts sits under Set up in the orchestration dashboard and does one thing — draw how your terminals work together, as an org chart or a pipeline. Start from the terminals you already have open, describe the team in plain words and let an AI draft it, or pick a ready-made shape; nothing runs until you press ▶ Start team. The Router became Defaults and moved to Advanced, where it now carries fallback chains — an ordered list of backups for any type of work, with the triggers you choose and the hop written down when it happens. Plus a Resource Manager behind the status-bar memory reading, Cursor terminals that read as chat, and Ctrl+click file paths that resolve against the terminal's own working directory.

The Charts canvas in Agent Orchestration with a Hierarchy org chart — Boss & Planner on Claude managing Researcher on Codex and Builder on Grok — beside the presets palette and a green Start team button
The New chart dialog offering Hierarchy, Pipeline and Router (Defaults) as chart types, with Use AI and Do it yourself as the two ways to start

New Features

  • +Charts is now one library for every org chart and pipeline — it lives under Set up in the orchestration dashboard and does exactly one thing: draw the relationships between your terminals, who reports to whom and who hands work to whom. A New chart dialog picks the shape (Hierarchy, Pipeline, or the Router/Defaults table) and how you start it — Use AI or Do it yourself — instead of hiding the choice behind a mode switch.
  • +Start from the terminals you already have open — with two or more AI terminals running, Agent Input offers “Draw how these N terminals work together”, and Charts' Chain my open terminals card drops exactly those terminals into a pipeline, already assigned.
  • +Describe your team in plain language — a ✨ Describe with AI popup with a real text box, clickable example prompts and a three-step guide (Describe it → We draw a draft → You press ▶ Start team) drafts the chart for you. A Generate with chip picks which installed AI CLI does the drafting, and shares one preference with the Draw canvas. Drafting only replaces the canvas — ⌘Z brings the old chart back, and it never starts a team.
  • +Start screens that show your options instead of hiding them — a blank chart, ready-made org charts and pipelines, your saved charts, and an explicit Convert my pipeline / Convert my org chart card, so converting a draft is a choice you make rather than a gate you pass through. Opening a saved chart over unsaved edits now previews what is on the canvas first, with Save edits first, Cancel or Discard and load.
  • +One Start per screen — while the seat table is open, the confirm at its foot is the only Start button, so you're never choosing between two controls for the same act.
  • +The Router is now Defaults, moved to Advanced — it is consulted at exactly two moments: work you never named an agent for, and a run that failed. Naming an agent always wins.
  • +Fallback chains — give any type of work an ordered list of backups, choose which failures trigger them (launch error, provider unavailable, rate limit, quota exhausted, no response), and decide whether the chain advances on its own or waits for you. Active teams show the hop that actually happened — Codex rate-limited → Claude.
  • +Defaults fills itself in on first run — a starting table is seeded from the AI CLIs you really have installed, instead of an empty grid. After enough runs it can also point out a type of work that keeps failing on one engine and propose a change; it never edits your table by itself.
  • +A routing board you can see whole — work types on the left, the agent that takes them on the right, sized to the space it has.
  • +Resource Manager — see what's using your machine — click the memory reading in the status bar for a per-process breakdown: 1DevTool's own processes (Main, Renderer, GPU) and every terminal, named the way its tab is. Live CPU sparklines run on each row while the panel is open, system memory, swap and CPU load sit underneath, and Clear terminal buffers releases cached scrollback so you can watch the number come down.

Improvements

  • ~Cursor and Grok now read as chat — a Cursor terminal shows its parsed transcript in Conversation and Reader Mode instead of a raw screen scrape, its sessions appear in Resume AI Sessions, and it resumes after a restart like every other agent.
  • ~Multiple-choice questions with mockups keep their shape — when an agent's question carries side-by-side ASCII previews, the transcript draws each option's mockup in its own monospace box with the columns intact, instead of a wrapped blob of JSON.
  • ~Ctrl+click a file path an agent printed and get the right file — paths now resolve against the terminal's own working directory (a git worktree included) before the project root, and only open when the file is really there.
  • ~Terminal mentions look like the terminals they name — an @terminal chip carries that agent's logo, and clicking it focuses that terminal.
  • ~Drag a search result straight into a prompt — file-search results in the explorer are draggable now, so a match can be dropped onto Agent Input as an @mention just like a file from the tree.
  • ~A real New terminal button — the tab strip's bare “+” is now a labelled control showing your actual shortcut, including one you remapped yourself.
  • ~Cursor and Cline own their quota pill — a focused Cursor or Cline terminal shows its own state and a link to that provider's usage page, instead of borrowing another provider's percentage.
  • ~Projects are recognisable everywhere — the project switcher draws each project's avatar, colour and emoji the way the sidebar does, and the same chip now marks Tasks cards, the agent roster, and assignment targets, so a task from another project reads at a glance.
  • ~Route & run shows its work first — batch assignment, single Route & run, and Agent Input's Orchestrate all open the same preview of which agent each task will go to before anything starts.

Bug Fixes

  • -Fixed multi-line prompts reaching Windows AI CLIs as their first line only — Describe with AI, the Draw canvas, diagram generation, and team delegation now deliver the whole prompt to Cursor, Pi and the other CLIs on Windows instead of one sentence.
  • -Fixed Conversation stuck on “Waiting for session transcript…” — a pane opened before the app finished loading its settings now retries and recovers on its own, instead of sitting on an empty chat while the terminal plainly shows output.
  • -Fixed Cursor's blinking caret sitting on the footer — the caret now appears in the Cursor composer, where you're actually typing.
  • -Fixed an approved plan leaving the run parked — approving a plan or a question now wakes the waiting agent, instead of going quiet until you type into its terminal yourself.
  • -Fixed broken screenshots in Software Update's release notes — release-note images load again instead of showing a broken-image icon.
  • -Fixed the routing board needing a horizontal scrollbar — with the rail, palette and inspector all open, the board fits the room it has, and its legend no longer scrolls away with it.

v1.66.0

August 22, 2026

Conversation — answer the agent without leaving the chat

Conversation now answers back. When a running agent asks one of those multiple-choice questions, Conversation shows it as an interactive card above the composer — pick an option, type a free-text answer, or check several at once — instead of sending you to Terminal. Team hand-offs read as clean rows too: who messaged whom, both agent logos, and a live delivery status. And a phone that lost its Remote connection can sign back in with a passcode instead of re-scanning a QR code. Plus a Builder that names the chart you're editing, hand-drawn Mermaid diagrams, clearer Agent Input setup switches, and fixes for Windows Codex pastes and frozen agent terminals.

Conversation view showing an agent's multiple-choice question as an interactive card above the composer, with numbered options like Review the diff and Run the tests and an Answer in Terminal button
The Conversation answer card in multi-select mode asking which surfaces matter, with checkboxes for Core engine, Desktop backend, Frontend and CI/build and a Next button

New Features

  • +Reply to the agent right in Conversation — when a running agent asks a multiple-choice question, Conversation shows it as an interactive card above the composer instead of only telling you to switch to Terminal. Pick a numbered option, type a free-text answer in the Type something row, or check several at once and press Next; a final tab lists what you chose before you Submit answers. Every choice becomes the keystrokes the agent expects, so the terminal stays the source of truth.
  • +Watch your team talk to each other — when agents in a team, hierarchy, or pipeline message one another, Conversation shows a tidy row for each hand-off instead of the raw [1devtool] wrapper and CLI command. A link-message card names who sent it and who it's for with both agent logos, folds the reply instructions away, and notice rows cover the small stuff (Linked to Codex, You are Backend lead). Outbound rows show what happened — Delivered · awaiting reply, Reply received, Waiting for your approval, or Delivery failed — with a line that animates while a message is still in flight.
  • +Reconnect your phone with a passcode — a phone that lost its Remote connection (expired pairing, cleared browser data, a different browser, or a device you revoked by mistake) can sign in again with a passcode instead of scanning a fresh QR code. Let 1DevTool generate a short one-time code or set your own, choose the permission every passcode-paired phone receives, and manage it all in Settings → Remote — off by default, changeable even while Remote is stopped.

Improvements

  • ~The orchestration Builder tells you which chart you're editing — the draft now carries a real name and two plain-language chips (saved and applied), so "is this saved?" and "is this what ▶ Start team runs?" are answered at a glance. Rename in place, Save back to the preset you opened or Save as… a copy, and switch between Hierarchy and Pipeline without threatening to convert your chart — each shape keeps its own draft, and converting is one explicit choice you can undo.
  • ~Diagrams look like real Mermaid charts — the graph editor now draws soft lavender nodes with the signature hand-drawn hatch fill, purple borders, and gently curved connectors, with a cleaner dark mode that keeps decision diamonds legible.
  • ~The setup switches in Agent Input show they're working — clicking Routing, Hierarchy, or Pipeline now reads Choose… or Starting… with a visible track, tells you what it's waiting for, and offers a Keep Off way out, instead of looking like nothing happened.
  • ~A finished worker's report reaches a busy manager on its own — when a worker reports back to a Claude director that's mid-task, the report now waits and delivers the moment that turn ends, instead of parking until you press "Submit now".

Bug Fixes

  • -Fixed large pastes into Codex on Windows getting cut off or sent early — a big block of text now lands in the Codex composer in full and waits for you to send it.
  • -Fixed Windows agent terminals freezing after "continue" — a terminal that reported its transport became uncertain and stopped accepting input now recovers, holds your keystrokes safely, and offers to restart a stuck seat instead of silently swallowing what you type.
  • -Fixed the orchestration Router not listing a project's own skills — a role routed to Codex (or any agent) now sees the skills installed in your project's shared folder, and a project skill correctly overrides a global one of the same name.

v1.65.0

August 21, 2026

Browser profiles — a separate login for every tab

Browser profiles give any tab its own identity — separate cookies, storage, and history — so Work, Staging, and a throwaway test account never bleed into each other, and an agent driving a tab works as the account you signed in. Set a default profile for new tabs or keep everything on one Shared session, switch a tab's login from a per-tab picker in the toolbar, tell sites to Look like Chrome (so sign-ins and checkouts behave) or use 1DevTool's native identity, and browse or wipe each profile's history — all in Settings → Browser. A new 1AIVault source in the Memory panel brings your vault into the workspace: browse, search and filter every saved memory and hand any entry to an AI agent as context, read-only by design. Conversation also gains a needs-your-answer banner when an agent stops on a question, shows your message the instant you send it, marks Reader Mode snapshots as captured-when-opened, keeps long transcripts smooth with Load earlier, and lets you resize the composer and jump back through your own turns. The terminal pane header is regrouped into Output · Render · Window clusters with a clearer Conversation Terminal toggle.

1DevTool Settings → Browser → Profiles tab with a Shared session card beside the mode (default) and khoa-2 separate-login profiles, each keeping its own cookies and history, plus an Add profile tile
Per-tab profile picker open in the browser toolbar, listing Shared, mode (checked) and khoa-2 under 'Open this tab as', so one tab can switch to a different signed-in account

New Features

  • +Give any browser tab its own identity — each profile keeps its cookies, storage, and history separate, so Work, Staging, and a throwaway test account never bleed into each other. Agents inherit the profile too: a tab driven by browser automation uses the same signed-in session you see. Set a default profile for new tabs, or leave everything on one Shared session — a single cookie jar for the whole app.
  • +Switch a tab's login from the toolbar — a per-tab profile picker moves that tab onto another profile, so you can be signed into the same site as two different accounts side by side.
  • +Tell sites how the browser should look — choose Look like Chrome (recommended, so sign-ins and checkouts work normally) or 1DevTool's native identity for the occasional login page that needs it.
  • +Browse and clear a profile's history — each profile remembers where it has been, and you can remove entries or wipe a profile and its data. Add, rename, set default, and delete profiles in Settings → Browser.
  • +1AIVault memory in the workspace — a new 1AIVault source in the Memory panel lists everything saved in your vault alongside your local agent memory files. Search and filter it without leaving 1DevTool, and hand any entry to an AI agent as context or copy it to paste elsewhere. Read-only by design — 1DevTool never changes your vault, and points you to 1AIVault when it isn't installed.

Improvements

  • ~Conversation tells you when the agent is waiting on you — a running agent that stops on a question, confirmation, or picker raises a The agent needs your answer banner so you don't miss it while reading the chat.
  • ~Your message appears the moment you send it — a prompt typed into the Conversation composer shows as your row right away, before the agent writes it into the transcript, and reconciles once the real turn lands.
  • ~Reader Mode says when it's showing an old snapshot — while the agent is still working, Reader marks its view as captured-when-opened and offers one click to load the latest, instead of looking frozen.
  • ~Long transcripts stay smooth — Conversation keeps only the most recent messages mounted, with Load earlier for the rest, so big sessions scroll without lag.
  • ~Resize the Conversation composer — drag its top edge for more room to write; the height is remembered per terminal. A quick control also walks you back through your own turns in a long transcript.
  • ~Cleaner terminal pane header — pane controls are grouped into Output · Render · Window clusters with unified, round-capped icons and a clearer Conversation / Terminal toggle. Switching AI panes shows chat instantly, since recently viewed conversations stay warm.

Bug Fixes

  • -Fixed Conversation staying blank on a session's first turn — a brand-new session, a /clear, or a conversation you just followed with /resume now appears immediately instead of waiting out a refresh cycle.
  • -Fixed popping out a terminal that Reader Mode was showing — the pop-out is now blocked for a docked terminal instead of tearing down the live view underneath it.
  • -Fixed a stray divider line on plain shell panes — a non-AI terminal no longer leaves a stranded hairline at the edge of its header.