Skip to content

A calm interface for AI coding agents

See every session at a glance.

Click in only when needed.

Return to calm.

Download
GitHub
nextDialog
Working
Vibe Coding Academy
~/dev/vca
Working
NextDialog
~/dev/nextdialog
Idle
Data Pipeline
~/dev/pipeline
+
NextDialog~/dev/nextdialog
TuneCatch me up+ Terminal
NextDialog
Terminal 2
+
Edited app/components/AppBackground.tsx
Edited app/utils/settings.ts
Running npm run build...
⚠ Build failed: 2 errors found
src/settings.ts:42 — Property ‘gradient’ does not exist
src/settings.ts:58 — Expected 2 arguments, got 3
Claude is waiting for your input
Developer App~/dev/app
TuneCatch me up+ Terminal
Developer App
Terminal 1
+
Claude Code v2.1.89
Welcome back!
▐▛███▜▌
▜█████▛▘
▘▘ ▝▝
Opus 4.6 (1M context) with medium effort · Claude Max
~/dev/app
Recent activity
No recent activity
What's new
Increased default max output tokens to 64k...
Added allowRead sandbox filesystem setting...
/release-notes for more
▸▸ bypass permissions on (shift+tab to cycle)
medium · /effort
Scroll to explore

You run multiple agents now. That’s a good thing.

Five agents. Five contexts. Five places to check.

One calm interface to see them all.

“It’s not so much about deep work, it’s about how good I am at context switching and jumping across multiple different contexts very quickly.”

Your agents, at a glance

nextDialog
Working
Vibe Coding Academy
~/dev/vca
Working
NextDialog
~/dev/nextdialog
Idle
Data Pipeline
~/dev/pipeline
NextDialog~/dev/nextdialog
TuneCatch me up+ Terminal
NextDialog
Terminal 2
+
Edited app/components/AppBackground.tsx
Edited app/utils/settings.ts
Running npm run build...
⚠ Build failed: 2 errors found
src/settings.ts:42 — Property ‘gradient’ does not exist
src/settings.ts:58 — Expected 2 arguments, got 3
Claude is waiting for your input
Developer App~/dev/app
TuneCatch me up+ Terminal
Developer App
Terminal 1
+
Claude Code v2.1.89
Welcome back!
▐▛███▜▌
▜█████▛▘
▘▘ ▝▝
Opus 4.6 (1M context) with medium effort · Claude Max
~/dev/app
Recent activity
No recent activity
What's new
Increased default max output tokens to 64k...
Added allowRead sandbox filesystem setting...
/release-notes for more
▸▸ bypass permissions on (shift+tab to cycle)
medium · /effort
See everything at a glance
Click in only when needed
Return to calm when you're done

Built for focus, not features

👁️

Glanceable Status

See every session’s state in two seconds. Color-coded dots, no reading required.

🎹

Session Tuning

Per-session hooks, permissions, startup commands, and reusable profiles. Each agent gets exactly the config it needs.

🎯

Priority Queue

“2 sessions need you” — click, resolve, return to calm.

⏱️

Catch Me Up

Stepped away? One click shows every edit, search, and build since you left.

Any Agent

Claude Code, Gemini CLI, Cursor Agent, OpenClaw, Terminal — anything terminal-based.

📈

Session Timeline

Append-only ledger of every action. See what happened, when, and why.

Works with every agent you use

Built-in support. Custom configs. Bring your own tools.

Claude Code
Codex CLI
Aider
Gemini CLI
OpenClaw
Plain Shell
+Create your own

+ Add custom type in 60 seconds with the guided Learn mode

Create a custom agent in 60 seconds

1
Name it
Name:OpenClaw
Icon:Color:
2
Set the command
$ openclaw tui
3
Learn patterns

NextDialog watches your terminal and suggests status patterns automatically.

4
Done
OpenClaw
idle

“Technology should inform but not demand our focus or attention.”

— Mark Weiser & John Seely Brown
Xerox PARC, 1995

NextDialog optimizes for you.

When your agents are working, the best thing you can do is trust that you’ll know when they need you.

Lightweight by design

~0MB
Binary
~0MB
RAM
<0ms
Launch

Built with Tauri 2 + Rust. Native. Not Electron.

Less managing.
More conducting.

Download
View on GitHub

macOS · Windows · Linux