Popups & Modals

Every modal in Cerevisor, what opens it, what's inside, and when it appears automatically.

Cerevisor has 30+ popups. Most you'll never see unless you trigger them; a handful appear automatically (provider setup, trust profile, update prompts, crash recovery). This page is the master list.

Use Ctrl/Cmd+F to find what you're looking for.

Configuration popups

Agent Config

The full editor for a single agent.

  • Opens: Right-click an agent → Configure, or click the agent's role badge / name, or select the agent and press Enter.
  • Contents: Name, role badge, role description (textarea), instructions (textarea), output definition (prose + optional structured fields), skills (ability slots), input files, output files, provider + model picker, search provider, subagent type, allowed tools, fallback strategy, fallback instructions, context notes.
  • Save behavior: Most fields commit immediately. Long-form text fields (instructions, role description, output prose) debounce 180 ms after you stop typing.
  • Close: Click outside, click the X, or press Esc. All pending edits are flushed on close.
  • See Guides → Building Workflows → Add and configure agents for the deep walkthrough.

Connection Config

Editor for a single connection between two agents.

  • Opens: Double-click a connection line, or right-click → Configure Connection.
  • Contents: Flow type (Always / Conditional / Approval Gate / Loop), data description, handoff format (Summary / Structured JSON / File Reference / Full Output), handoff schema (for structured-json), condition (for conditional), approval prompt (for approval-gate), loop condition + max iterations (for loop), conditional edge label, else-branch flag.

Loop Config

Per-agent loop configuration (run this agent N times, with this exit condition).

  • Opens: Click the Loop icon at the bottom-right of an agent card.
  • Contents: Loop type, exit condition (natural-language or structured), max iterations (default 3), per-iteration prompt template.

Schedule Config

Configure a cron-style schedule for a workflow.

  • Opens: Click the Schedule icon at the bottom-right of an agent card (schedules just that agent), or the schedule button on the workflow control bar (schedules the whole workflow).
  • Contents: Cron expression (with a human-readable preview), timezone, run mode (full workflow or just this agent), and next-fire-time display.

Workflow Input Config

Defines what input the workflow accepts when run.

  • Opens: Settings → Input tab, or click the input-label pill in the status bar.
  • Contents: Input type (None / User Message / File / Folder — Webhook is shown but marked "Coming soon"), the input value (message text, file, or folder), and the entry agent that receives the input.

Run-time popups

These appear during a workflow run.

Execution Modal

Live status of an active run.

  • Opens: Automatically when you click Run.
  • Contents: Per-agent status cards (waiting / running / completed / failed), streaming output preview, tool calls, current wave indicator, run cost so far, run duration.
  • Buttons: Stop, Pause, Resume, View full output for any agent.
  • Survives Escape: No: Esc closes it, but the run continues in the background. Reopen via the status bar.

Permission Modal

Approval request for a tool call.

  • Opens: In Oversight mode, before every tool call. In Auto mode, before a consequential action (deleting files, sending anything, a tool from an untrusted integration).
  • Contents: Tool name, target path or command, a Routine/Consequential badge, the agent that's asking, and a plain-language line explaining why the prompt appeared.
  • Buttons: Approve once (this call only), Always allow [tool] (auto-approves that kind of action from that tool for the rest of this run — approving routine commands does not blanket-approve deletes), Approve all N pending (when multiple requests for the same tool are queued), Deny (tool returns an error to the agent).

Approval Gate Modal

A connection between two agents is marked as an Approval Gate: the workflow pauses for your sign-off.

  • Opens: When a workflow reaches an approval-gate edge.
  • Contents: The upstream agent's output preview, the approval prompt you wrote on the connection, the downstream agent that's waiting.
  • Buttons: Approve (continue), Reject (stops the run, marks the connection failed).

User Question Modal

An agent called the ask_user tool to ask you something.

  • Opens: Automatically when an agent invokes ask_user.
  • Contents: The agent's question, optional context, optional list of suggested answer choices.
  • Buttons: Pick a suggested choice, or type a free-form answer and submit. Cancel stops the agent.

Interrupted Recovery Modal ("Recovered interrupted work")

Cerevisor crashed or was force-closed during a run.

  • Opens: Automatically on next launch if interrupted runs are found. All saved progress is restored to the run view; interrupted agents are left stopped and won't spend or act until you resume one.
  • Contents: The workflow name, when the run started, how many turns and how much cost were preserved, and the list of interrupted agents.
  • Buttons: Resume agent (re-runs just that agent; only enabled when its workflow is the one currently open), View recovered work (opens the run view on the recovered session), Discard (deletes the session). Clicking outside dismisses without deleting — the run's history stays available in Analytics.

Agent Plan Modal

A preview of an agent's generated plan before it executes.

  • Opens: When an agent's plan-mode produces output you've asked to review.
  • Contents: The plan steps, tool calls the agent will make.
  • Buttons: Approve (run the plan), Deny (stops the agent run with a plan-denied error).

Setup popups (first run / one-time)

Welcome Intro Modal

First-launch orientation.

  • Opens: Automatically on the first launch after install. Once dismissed, won't reappear unless re-enabled in Settings → Learning Experience.
  • Contents: A short tour of what Cerevisor does, with slides you can advance or dismiss.

Provider Setup Modal

Configure a model provider.

  • Opens: Automatically on first launch (after Welcome) if no provider is configured. Also opens from Settings → Providers → + Add provider.
  • Contents: Provider type picker (Anthropic / Gemini / Claude Agent SDK / OpenAI-compatible / Codex CLI / Cursor Agent / Antigravity), then a provider-specific configuration form.
  • See Connecting a provider for the full walk-through.

Approval Mode Setup Modal

Choose how often Cerevisor pauses for approval.

  • Opens: Automatically on first launch after Provider Setup. Also re-openable from Settings → Permissions.
  • Contents: Three modes: Oversight (approve each action), Auto (recommended — runs by itself, asks before consequential actions like deletes and sends), Hands-free (never asks; requires a one-time confirmation the first time you pick it). Capability limits in Settings → Permissions apply in every mode, including Hands-free.

License Setup Modal

Activate a license or start a trial.

  • Opens: Automatically on first launch and on revalidation failures.
  • Contents: License key entry, trial countdown, link to upgrade.

Loop & Enhance Setup Modal

First-run configuration for the Loop & Enhance workflow improvement system.

  • Opens: Automatically the first time you enable Loop & Enhance on a workflow. Subsequent uses go through the Loop & Enhance tab of the run view.
  • Contents: Max rounds, total cost cap, convergence mode, pause-between-rounds toggle.

Browsing and viewer popups

Skill Viewer

Read-only rendered Markdown viewer for a skill's documentation.

  • Opens: Double-click a skill in the Skills panel, or double-click a skill hexagon on an agent.
  • Contents: The skill's SKILL.md rendered with syntax highlighting; metadata header (name, description, tags).
  • Buttons: Edit in Workshop (opens the Skill Workshop with this skill loaded; trial/paid only), Open in file explorer, Close.

File Viewer

Read-only rendered viewer for a markdown file attached to an agent.

  • Opens: Click an input/output file chip on an agent, or click a file in the Markdown panel.
  • Contents: Rendered Markdown with syntax highlighting and table support.
  • Buttons: Open in file explorer, Close.

Version History Modal

Browse past versions of the current workflow.

  • Opens: Click the version-history icon in the title bar.
  • Contents: List of saved snapshots with timestamps and a brief change summary.
  • Buttons: Preview (read-only), Restore (replaces current workflow), Delete snapshot.

Memory History Modal

Browse monthly snapshots of your memory documents.

  • Opens: From the Memory view → History tab.
  • Contents: Snapshot list with diff previews.
  • Buttons: Restore snapshot, Compare to current.

Cloud Sessions Modal

Manage Cursor cloud runs.

  • Opens: Title bar Cloud Sessions icon (only visible if you have any cloud runs).
  • Contents: Active runs, recent runs, with metadata (workflow, agent, started-at, status, agent ID, run ID).
  • Buttons: Reconnect, Cancel, Dismiss, Refetch artifacts, Open artifacts folder.

Agent Manager Popup

Browse, export, and import every agent in your current workflow.

  • Opens: From the manage button on the agent palette or the command bar.
  • Contents: List of agents with quick stats; export to template, import a template.

Skill Telemetry Popup

Per-skill invocation telemetry.

  • Opens: From the Skill Viewer's telemetry button (if any data exists).
  • Contents: Call count, success rate, average duration, recent invocations.

Workspace popups (World View)

Add Workflow Modal

Add a new workflow to the current workspace.

  • Opens: Click the "+" on the world view rail, or from Settings → New workflow.
  • Contents: Workflow name, folder path, template picker (blank or starter templates), initial position on the canvas.

Configure a link between two workflows.

  • Opens: Drag from one workflow frame to another in the world view.
  • Contents: Source agent (which agent in workflow A produces the data), target agent (which agent in workflow B receives it), data description, handoff format.

Embedded App Picker Modal

Pick a desktop app to embed as a canvas node (Windows only).

  • Opens: Canvas context menu → Embed app, or the Embedded Apps dock.
  • Contents: Grid of supported apps (Outlook, Excel, Word, Edge, Explorer, Notepad, Generic) with their icons.

Embed Data Selection Popup

Pick which slice of an embedded app's data to feed into an agent.

  • Opens: When you drag a connection from an embedded-app node to an agent.
  • Contents: Available feeds for that app (e.g. Outlook: emails, calendar, contacts).

Operator Panel

The home of the Operator — an assistant that works quietly on your workspace in the background while the app is open, within a budget you set.

  • Opens: Click the clock icon on the workspace rail (World View). The Operator is workspace-wide, so it lives on the rail rather than inside any single workflow.
  • Contents: A header (on/off, how far it may reach — read-only here; widening lives in Settings → Automation — when it next checks for work, today's spend, and a Stop button), a journal of everything it has done, your objectives (status, spend, and any question it's parked for you), and a review queue where everything it makes waits — with a change preview — for you to Accept or Discard.

Configure a multi-workflow pipeline run.

  • Opens: Title bar Pipeline icon, or Settings → Automation → Pipelines.
  • Contents: Ordered list of workflows to run, per-step input mapping, budget cap, share-skill toggle, share-permission toggle.

Help and meta

Shortcuts Overlay

The keyboard shortcuts cheat sheet.

  • Opens: Press ? or Ctrl+/, or click the keyboard icon in the title bar.
  • Contents: All shortcuts grouped by area. Same content as keyboard-shortcuts.md.

Update Ready Modal

Notifies you when a new version of Cerevisor has been downloaded.

  • Opens: Automatically when an update finishes downloading in the background.
  • Buttons: Restart now (applies update immediately), Later (applies on next manual restart).

Settings

The omnibus settings popup.

  • Opens: Title bar gear icon, or any link that says "open Settings".
  • Contents: 17 tabs. See Settings reference for the full breakdown of every tab.

Confirm modals

Cerevisor uses two confirm-modal variants. They have no titles in this list because they're not "named" popups; they're a UI primitive used wherever a destructive action needs sign-off.

  • Confirm Modal: Generic yes/no. Cancel + Confirm buttons. Appears for "Save before opening?", "Discard draft?", etc.
  • Destructive Confirm Modal: For irreversible actions like deleting a wave full of agents. Requires you to type a confirmation phrase. Appears for "Delete wave (3 agents)", "Clear workflow", "Delete provider used by N agents".

Toast notifications (not strictly popups)

Cerevisor uses non-modal toast notifications for transient feedback:

  • Save successful / Save failed.
  • Skill added to [agent name].
  • Provider added / Provider removed.
  • Workflow tip dismissed.
  • File scan complete.

Toasts auto-dismiss after a few seconds. Click the X on a toast to dismiss early, or click the toast body to act on it (some toasts have actions, like "Provider added, set as default?").

Back to docs