One terminal is not enough anymore
Coding agents got good enough to work in parallel. Our tools didn’t. Here is the idea behind PaneScale.
Notes on working with coding agents.
Practical, tested advice for running agents in parallel, verifying what they produce and keeping them affordable — most of it useful with or without PaneScale.
Five ways an agent can fake a green build
Agents rarely cheat on purpose, but under pressure to make tests pass they find shortcuts. Here is what to look for — and what PaneScale flags automatically.
Checkpoints that never touch your branch
How PaneScale snapshots a working copy byte for byte without moving HEAD, the index or any branch — and why rewinding refuses to overwrite your edits.
Write tasks an agent can finish
The difference between a task that runs overnight and one that wanders is usually a single sentence — the finish line. A short template and examples.
Bypass mode: a decision table
Skipping an agent’s permission prompts saves hours — and occasionally costs a lot more. When it is reasonable, and what to put in place first.
Why the second pane opens instantly
Starting a coding CLI takes seconds. Pre-warmed sessions hide that wait — here is how we do it without eating your memory.
What a good morning report looks like
An overnight run is only as useful as what it leaves behind. The rules behind PaneScale’s morning report, and how to review one in ten minutes.
Measured, not estimated
Why the Token Guardian shows “not reported” instead of a guess — and how real budgets stop runs instead of turning numbers red.
A grid you can drag into any shape
Design notes on the resizable terminal grid in PaneScale 0.2 — seams, crossings, snapping and why the shape is saved per workspace.
Memory is context, not commands
Teammates in PaneScale remember. Here is why everything they remember is passed as context and never as instructions — and why that matters for safety.
Shipping local voice with whisper.cpp
How PaneScale does dictation and voice commands without sending audio anywhere — model pinning, temporary recordings and commands we deliberately left out.
Designing an MCP server agents can’t abuse
PaneScale’s MCP bridge lets any agent work with your task board. These are the decisions that keep it from becoming a way around your approvals.
Every run should leave a receipt
When agents work in multi-step workflows, “it worked” isn’t enough. What a run record should contain, and how PaneScale’s workflow runs provide it.
Product videos as code
Why Video Studio describes every scene as a small canvas function — and the determinism check that keeps preview and export identical.
Never present a guess as exact
Point & Fix has to find the source of whatever you click. Sometimes it knows; sometimes it guesses. Saying which is which is the whole design.
Git worktrees on Windows, without the pain
Worktrees are the best way to isolate parallel agents. On Windows they come with a few traps — path length, dependencies and cleanup. Here is how to avoid them.
What the next agent needs to know
When one agent gets stuck and another takes over, the handoff decides whether the second attempt is better or just different. Anatomy of a continuation brief.