Ghost Architect™ v7.0.0: Ghost Open is the trial

This post covers what shipped, why we removed the trial flow, the new cost transparency in Question mode, and what existing users need to know.

Why we removed the trial flow

Most dev tools gate the good parts behind a signup, a credit card on file, or a 14-day timer that starts the moment you install. That model trains users to evaluate quickly, often before they have done anything real with the tool. It also creates an adversarial conversion path: the user is racing a clock they did not ask for.

Ghost Architect™ does not work like that. The conversion path is honest. Run Question mode unlimited. Run Recon mode unlimited. Burn through four free deep reports across Points of Interest, Blast Radius, Conflict Detection, and Prompt Triage on whatever codebase you want. By the time you hit the paywall, you know exactly what the tool does, on your code, with your stack, with your API key. No signup. No timer. No friction.

Inheritance Audit stays Pro and above. That is the deal-grade audit, and that is where the buyer conversation starts.

Cost transparency for Question mode

Question mode now shows the cost estimate before you run the call and the actual cost after. Two numbers. No surprises.

Most AI dev tools hide the bill. They run the call, return the answer, and let the meter run silently in the background. You find out at the end of the month. Ghost Architect™ shows you what every Question is about to cost in API tokens, lets you confirm or cancel, then prints the real cost when the result lands.

This is the same cost pre-flight that already runs on every paid scan mode. Now it runs on the free one too. The math should be visible. The bill should not be a mystery.

Profile feature properly tier-gated

Ghost Partner™ profile support, the --profile, --create-profile, and --list-profiles flags, and the in-CLI profile menu are now properly restricted to Pro, Team, and Enterprise. The Open tier sees a clear upgrade message instead of a broken flag.

This was a cleanup pass. The feature shipped in v5.0 but the tier gate had drifted in a few code paths. v7.0.0 makes the gate consistent everywhere.

Model defaults updated

Default model lineup is now Sonnet 4.6 and Opus 4.7. In practice this drops API cost for most scan modes. The cost preflight reflects the new pricing automatically.

README rewritten

The README on GitHub and on npm is end-to-end clearer, more accurate, and better structured. If the install instructions confused you in v6, read them again in v7.

For existing users

3,933 people are already running Ghost Architect™. v7.0.0 installs automatically on the next npm update -g ghost-architect-open. Your config persists. Your scan counts persist. Nothing breaks.

If you are on a paid tier, your subscription continues unchanged. The new cost transparency and the new model defaults apply immediately.

Install Ghost Open

npm install -g ghost-architect-open

Run it on something. See what it finds.

EJ Wisner, founder of Ghost Platform™