Repository navigation
Releases
Chord v1.0.0-beta.7
The seventh pre-release of the next major.
A session that starts others spends less. The rest of the work on what a long orchestration costs.
- Reports arrive together. When several started sessions finish close to each other, their starter is told in one message and not once for each. While others are still at work it is told shortly after the last one rests, within two minutes of the first. A starter whose conversation has gone cold is told once all of them are done, or within twenty minutes. A session that fails or asks something is told at once, and a
waitstill answers at once. - A started session keeps a smaller context. It compacts at 150,000 tokens, so a long piece of work does not carry everything it has read into every later step.
- A started session runs one model below its starter when the start names none. A start that names a model gets that model, and a workflow step's own default still decides for that step.
- A local usage log. Each session's tokens per turn and its context's size are kept in a small file per session on your machine, bounded in size and shown to nobody, for tuning these thresholds.
In the conversation. A read of the session's notes is not drawn as work, like the saving of one, so an answer followed by it stays the last thing in its run.
Resume on reset. A Claude Code conversation on an account with a five-hour usage window has an auto-resume chip beside the model, the effort and the mode, off by default. With it on, a turn that window cut short is waited for: the row and the status bar say when it resumes, and a minute after the reset the session goes on by itself, as a fresh conversation from its notes. It never resumes a turn you stopped or one a weekly limit ended, it tries once, and the wait holds only while the app runs. Turning it off ends the wait.
The earlier name is gone underneath too. The settings folder is ~/.chord, with a link left at ~/.agent-workbench; the tool server and its tools are chord; the daemon is chord-remote; the variables are CHORD_*. The app moves an existing install over at first run, and keeps using the old folder if the move fails. A paired machine's daemon must be upgraded with the app.
An answer that never appeared (#196). The agent had written it as thinking, which is not shown, and not as a message. An agent in a conversation is now told that whatever you should read goes in a message's text.
After installing. Every session's first message after the upgrade reads its conversation once more, since the tools changed.
Installing. brew upgrade on the beta channel, or brew install --cask elva-labs/elva/chord@beta. A paired machine needs the daemon at the same version: brew install elva-labs/elva/chord-remote-beta there, or the install script with CHORD_VERSION=v1.0.0-beta.7.
Chord 1.0.0-beta.6
The sixth pre-release of the next major.
A long conversation is cheap to come back to. Coming back to a conversation after a pause read all of it again at full price, which for a long one could take a large share of a usage allowance in one message. Every session now keeps a brief and notes as it works: what was decided and why, what was found, what you asked for, what is open. When you return to a Claude Code conversation that has gone cold and grown large, it goes on from those, with its last exchanges word for word, and not from the whole conversation read again. It is the same session to you: the same row and name, the earlier messages still shown, under a line that says it went on from its notes. The agent reads the notes that match when it lacks context, and can search the old conversation for a detail. The row's menu shows the notes, and resumes the old conversation whole, saying what that reads. A conversation from before this release has no notes yet and goes on from its first message and last exchanges. Codex conversations resume whole as before.
The app tidies up after sessions that start others. A session another session started is put away by the app five minutes or more after its work is over, its starter told and nothing of it open, and never while it is unread or holds a draft. A line from its starter brings it back.
Less at the top of the window. A session's state, its folder and its branch are in the status bar. The account stands there too, as account: <name>, when more than one is configured, and the sessions pane has no foot. The Agent / Editor switch sits in the right corner. The changes header no longer counts the files.
Panes move smoothly. Closing and opening the viewer, folding the sessions pane and opening the terminal run at the display's rate from the first frame, the terminal's first opening as its second.
In the conversation. The composer has a chip for the effort between the model and the mode. A fenced block of prose wraps, while code and text laid out in columns keep their lines. Such a block, the card for a started session, the "Worked for" line and the app's notices end where the prose ends, and headings and blocks keep even room between them. A session is not named after a slash command, and /rename, /effort and /model reach the row and the chips.
Also. A presented Markdown document uses the viewer's width. A double-click in the empty editor starts a file. A project's past sessions come last, behind a fold that closes again. In a workflow step's list of models the one used when none is named is marked "default", with "make default" on the others.
After installing. Every session's first message after the upgrade reads its conversation once more, since the tools changed.
Installing. brew upgrade on the beta channel, or brew install --cask elva-labs/elva/chord@beta. A paired machine needs the daemon at the same version: brew upgrade agent-workbench-remote-beta there, or the install script with AGENT_WORKBENCH_VERSION=v1.0.0-beta.6.
Chord 1.0.0-beta.5
The fifth pre-release of the next major, and the first as Chord.
Agent Workbench is now Chord. A new name and a new icon, a circle cut by one chord. Your settings, projects, sessions and sign-ins carry over: the app keeps its identity underneath, and its settings folder and the tools agents see keep their names for now.
- macOS: the app is
Chord.app, so pin it to the Dock again. If you installed from the disk image, quit and deleteAgent Workbench.appfirst. - Windows: the setup replaces the earlier install and asks before closing a running app. Pin Chord again.
- Linux: the package is
chordand replacesagent-workbench. Add it to your favourites again.
Go back to an earlier message. In a conversation, Esc Esc with the composer empty and the agent at rest opens a list of your earlier messages. Pick one and the conversation goes back to just before it, with the message in the composer to edit and send again. Claude Code can put the files back as well, and the choice names them; Codex goes back in the conversation alone. A single Esc during a turn still only stops the turn. It is also in the palette and as "Go back to here" on each of your messages.
Up takes a waiting message back. With the composer empty, Up takes your newest waiting message back to edit, and Up again the one before.
A new session starts as you left the last. The model, effort and permission mode you last picked for an agent are what its next new session starts with. Plan and the unguarded modes are never carried over. Settings, Sessions, shows what is kept and clears it.
Unread stays unread. A row that was unread when you quit comes back unread with its line, whether it went unread by itself or you marked it.
Compact mode. The pane comes in however fast you reach the left edge, also beside the status bar and against the screen's edge, and no longer during a drag. The mark at the edge that read as a drag handle is gone; the status bar's count says what waits.
A session's state is said once. The status bar no longer carries the word "working", which was the window's mode and read as the session's state. The header, the row and its dot now agree, an idle session reads "idle" in a terminal as in a conversation, and a session at rest whose sessions work says so. The status bar's text and the agent header's title are centred.
A session moved to a conversation shows its history. Under a named account a session moved from a terminal could come up empty though the agent had the whole conversation. Where earlier messages cannot be read the thread says so and why, and keeps saying it.
For sessions that start others. A start can name your own login as Default. A start that waits on your answer is answered in time and told once how it went, so the same work is never started twice.
Also. An embedded browser tab could be left showing its address where it had a title, or a greyed-out Back with somewhere to go; the newest state now always wins. The daemon installs on systems that keep only a shell in /bin.
To check after installing. Click a notification and see that it lands on the question. In compact mode, flick the pointer to the left edge, also on a second screen.
Installing. brew upgrade moves an install of agent-workbench@beta over to chord@beta. A fresh install is brew install --cask elva-labs/elva/chord@beta. A paired machine needs the daemon at the same version: brew upgrade agent-workbench-remote-beta there, or the install script with AGENT_WORKBENCH_VERSION=v1.0.0-beta.5.
Agent Workbench v1.0.0-beta.4
The fourth pre-release of the next major.
The background shows the colours you pick. At full strength a background was near black with a faint tint. Strength now runs evenly from a calm tint to the colours as the pad shows them, in steps of 1, each hue rising to where it reads clean. Nothing is held back to protect text any more: text turns to black or white where it has to, and a quiet line under the controls says when it gets hard to read. A background kept near the top of the old range becomes the colours themselves; the old look of 100% is now at 16%. A Cards share the app used to hold back shows whole. Dragging a colour is smooth: about 19 ms a frame where it took 43.
Compact mode. A setting under Appearance hides the sessions pane and gives the other panes the window's width. The pointer at the left edge slides the pane in over the content, and it slides away after the pointer leaves it. The chord for the sessions pane (⌘1) brings it in with the keyboard in it, and the same chord or Escape sends it back. A thin mark at the edge shows while a session waits for you.
A notification when something waits for you. Every question that waits on you sends one system notification while the window is behind: a permission or a question from a session, one of a session another started, the app's own question before a session starts another, and an agent's change to the settings. A click brings the window forward at the question, and the notification is taken away once answered. It says no more than the row would: a command's program, never the whole command.
A started session begins in its starter's mode. A session started by one in Auto began in Default under some accounts and asked for everything. It begins in its starter's permission mode now, never a wider one, and a start in the caller's own project, on its agent and account, is no longer put to you. A setting under Sessions asks before every start if you prefer. A session is told to stop the sessions it started once their work is taken up, and tools take the first eight characters of a session's id.
Mark as unread. A session row's menu starts with Mark as unread. The row turns as one that went unread by itself, with its last notice under the name, and stays so until you have left the session and come back, or write to it.
Switching between terminal and integrated applies to open sessions. Changing how an agent's sessions run moves the open ones too, on the same row with their conversation. A session at work, with a question or a draft, moves once it is at rest, and one that would lose a running command says when it moves.
In the conversation. An answer stays in sight while the agent waits on sessions it started, and the wait itself is not listed. The workbench's own tools read in plain words ("Waiting on the sessions it started"). Messages from the workbench or another session stand on the left, with sessions named by their row. The caret is the height of the text, and the last line keeps its room above the composer.
A presented page shows the files beside it. An HTML page an agent presents loads images, stylesheets, fonts and media from its own folder, and nothing else: no network, no file above the folder. A file that cannot be loaded leaves a dashed box and its reason under the page.
Also. A project's + starts nothing but the session you then pick. The focused card's accent line no longer runs across the sessions pane, and the folded column's rows are no longer dimmed at their end. A line that begins with ! is refused to a session in a terminal, so no session makes another run a shell command unasked.
To check after installing. Notifications were run on a build without a hand to click them: click a banner and see that it lands on the question. In compact mode, see that nothing sits under the window's traffic lights.
Installing. brew upgrade on the beta channel, or brew install --cask agent-workbench@beta from the elva-labs/elva tap after brew uninstall --cask agent-workbench. A paired machine needs the daemon at the same version: brew upgrade agent-workbench-remote-beta there, or the install script with AGENT_WORKBENCH_VERSION=v1.0.0-beta.4.
Agent Workbench v1.0.0-beta.3
The third pre-release of the next major. It closes what beta.2 listed as known and what turned up since.
An integrated session starts in the mode you set. A conversation started in the Default permission mode whatever the agent's own settings said, so it asked where a terminal session would not. It takes the agent's settings now, and a session another session starts does too. A row can be reopened the other way, as a conversation or in a terminal, and keeps the mode its chip last showed.
Send now, and take back. A message written while the agent works waits under the thread. The line under the last waiting message sends them now, and each waiting message of your own can be taken back into the composer. Claude Code takes the message into the running turn where it can and stops the turn for it otherwise; Codex steers it into the running turn. The waiting line stays above the composer.
A long conversation reaches its foot. In a conversation of several hundred turns the view could stop short of the last line, and switching between long conversations stalled the window. The foot is reachable by wheel, keys and the scroll bar, and a conversation that is not on screen costs no layout.
Settings and sliders answer at once. Dragging the Cards control or a colour repainted every terminal on each step. Terminals are recoloured only when their colours change, and opening the settings no longer stalls.
Usage figures stay fresh. The status bar's usage is fetched for each account in use about every five minutes and when a session of an account without figures comes to the front, and the last figures are kept across a restart with their age.
A session that starts others hears the truth. A started session was reported as stopped during any quiet spell of a long turn. A turn counts as ended only when the agent says so or you stop it. A turn an error ended is said for what it is, a usage limit with the time it lifts; a stop is told once; the list says which sessions are closed and which can be resumed; and a name given at start stays through a close and a resume.
Also. In a Codex session, in a terminal or integrated, ten of the workbench's own tools run without asking: the ones that only read, and the ones whose changes the workbench asks you about itself. Every other tool, the browser and the ones that start, stop or write to sessions among them, follows the session's approval policy. The background's gradient is drawn finely once it rests, without stair-steps. The sessions list's scroll bar lies over the pinned project headers.
Installing. brew upgrade on the beta channel, or brew install --cask agent-workbench@beta from the elva-labs/elva tap after brew uninstall --cask agent-workbench. A paired machine needs the daemon at the same version: brew upgrade agent-workbench-remote-beta there, or the install script with AGENT_WORKBENCH_VERSION=v1.0.0-beta.3.
Agent Workbench v1.0.0-beta.2
The second pre-release of the next major. It fixes the freezes of the first.
The window froze with a large changes tree. Each row of the file tree held an ordering number of its own, and the macOS webview walks the list once for every distinct number on each layout and paint, so a tree of a few thousand rows made any layout take seconds: typing in a terminal, opening the tree, opening settings. The rows hold no number now, and the same layout takes a few milliseconds.
See-through cards show. In dark the tinted ground was painted lighter than the cards, so the two ran together, and a small share moved a card by a level or two. The ground is darker with more colour, a card takes on the background's colour and keeps its own lightness, and the Cards control runs from 0 to 100%. A share saved in beta.1 reads four times higher on the new scale and looks stronger.
The conversation spans its pane. An integrated session's thread, composer and dock use the pane's width; code, diffs, output and tables take all of it, and prose wraps at a comfortable measure.
Colours that belong together. Adding a third colour to a background places the others close to the first, about 50 degrees either side, and the first, drawn larger, carries them as it is dragged. The ready-made backgrounds are retuned. A background saved before is drawn as it was.
A read session stays read. A session's row went unread again a minute after it stopped, and again after each recap or compaction Claude Code did while idle, since its idle notice was taken for news. It is marked now only for something new.
Also. The focused sessions pane carries its mark under its header, as the other panes carry theirs along their cards. The orchestrator project is retired: to direct work, pick a workflow, among them a built-in Direct, or ask any session to start others; a folder of old orchestrator sessions is listed as an ordinary project while it has history. Two faults found by running Codex against a real model are fixed: a failed turn was told twice, and a session named no effort when Codex set none.
Known in this pre-release. An integrated session starts in the Default permission mode whatever the agent's own settings say, so it asks where a terminal session would not; a long conversation can stop short of its foot when scrolled; a message written while the agent works cannot be sent at once. Fixes for all three are in review.
Installing. brew upgrade on the beta channel, or brew install --cask agent-workbench@beta from the elva-labs/elva tap after brew uninstall --cask agent-workbench. A paired machine needs the daemon at the same version: brew upgrade agent-workbench-remote-beta there, or the install script with AGENT_WORKBENCH_VERSION=v1.0.0-beta.2.
Agent Workbench v1.0.0-beta.1
The first pre-release of the next major. It runs on the settings and state of 0.8.0 and hands them back to it, so moving between the two loses nothing.
Sessions as conversations. An agent can run integrated: Claude Code and Codex are driven through their own structured protocols and the conversation is drawn natively, with your words in a bubble, the agent's steps folded into one line, permission requests docked over the composer, and your own login, skills, hooks and MCP servers as before. A session in a terminal stays a choice in Settings. A resumed conversation shows its earlier messages. The composer completes slash commands and skills with /, mentions a file with @, and attaches any file. Cmd F finds in the conversation, each part can be copied, and a long conversation stays fast.
Workflows. A workflow gives sessions fixed roles, an orchestrator with work and review steps, drawn and edited as a diagram. A new session picks one from a chip in the composer. The header shows where a run stands, started sessions hang under their starter with their subagents, a starter answers its sessions' requests from its own dock, any session that started others gets a board at the foot of the changes pane, and an idle starter is woken when a session it started stops or asks.
A new frame. Panes stand as cards on the window's ground with their headers in one band, and a divider is drawn only while it is dragged. The sessions pane splits active from inactive by itself and sorts by last used or name. Cmd P goes to a session or project and Shift Cmd P runs a command.
A background of your own. Appearance takes a colour or a gradient behind the chrome, with strength and grain, and the cards can let a share of it through. Text keeps its contrast whatever is chosen.
Around the edges. Shells and the processes list belong to the session they were opened beside. Cmd T opens a shell, Cmd W closes what is in front, and closing the window or quitting asks while sessions or a busy shell are running. Editor mode takes untitled files, named when first saved. Browser tabs belong to the session that opened them, and an agent can take a screenshot of its own tab while you are elsewhere.
Installing. brew install --cask agent-workbench@beta from the elva-labs/elva tap, after brew uninstall --cask agent-workbench if the stable app is installed; the two are the same app and cannot sit side by side. brew upgrade then follows the beta. A paired machine needs the daemon at the same version: brew install elva-labs/elva/agent-workbench-remote-beta there, or the install script with AGENT_WORKBENCH_VERSION=v1.0.0-beta.1. Windows gets the NSIS installer and Linux the .AppImage and the .deb; a pre-release has no .msi and no .rpm.
Known. A custom stylesheet written for 0.8 may need changes for the new frame. Going back to 0.8.0 with the .deb is a downgrade to dpkg and has to be asked for as one.
Agent Workbench v0.8.0
A command palette. Shift Cmd P (Cmd ; in the vim preset) puts a field over the workbench listing every session and every action the keys can do. A session is listed in the order the list by status has it, with its project, host, account, the orchestrator that started it and the line it left; an action is listed with its chord. What is typed narrows the list by every word, in any order. Up and Down move, Enter picks, Escape or the chord again puts it away and the keyboard goes back where it was. While the palette is up no other chord gets past it.
The next session that needs you. Cmd . goes to the next session waiting on you and puts the keyboard in it: those asking for a permission first, then the one that has waited longest, stepping round from the session on screen so pressing it again visits every one. With nothing else waiting the status bar says so.
Zoom per terminal. Cmd =, Cmd - and Cmd 0 make the terminal with the keyboard a pixel larger or smaller, 8 to 24, or 12 again: the panel's shell when the panel has the keyboard, the agent otherwise. Each terminal has its own size, and an agent session's comes back after a resume or a restart. The grid is measured again and the process told, as for a font change.
All five are ordinary keymap entries, rebindable in Settings → Keys. A keymap saved before them takes the preset's chords for them, never one the keymap already gives to something else. Switching sessions with the keyboard in a terminal no longer leaves the keyboard on the body.
Upgrading. The desktop only pairs with a daemon at its own version, so a paired machine needs 0.8.0 as well: brew upgrade agent-workbench-remote there, or the install script again. On macOS, brew upgrade --cask agent-workbench from the elva-labs/elva tap.
Agent Workbench v0.7.9
style: format the delegated stop test
Agent Workbench v0.7.8
chore: bump to 0.7.8