Prompt Studio

Prompt Studio came out of a friction I felt every time I wanted to try a prompt idea: write it in a text file, then switch to a browser tab running some playground, then copy-paste it over, then copy the result back if I wanted to keep it. None of that friction does any actual thinking for you — it's just overhead between having an idea and testing it.

So I built the same authoring and testing experience directly into the editor, and I mean that literally rather than as a marketing simplification: the extension's build scripts run nx build prompt-workbench and nx build prompt-workbench-api, then copy the resulting Angular and NestJS bundles straight into the extension's webview/ and sidecar/ folders. There's no second implementation to keep in sync — when I fix a bug or ship a feature in Prompt Workbench, Prompt Studio picks it up on the next build. On activation, the extension spawns that sidecar as a child process and talks to it over plain HTTP, the same way the SaaS app's Angular front end talks to its NestJS back end; the webview is that same front end, rendered inside a VS Codium panel instead of a browser tab. It also supports multiple AI providers through the shared router, so the same prompt can be tried against different models side by side.

The result is a shared domain model between two products instead of two: Prompt Studio and the standalone Prompt Workbench web app aren't parity projects drifting apart over time — they're one codebase wearing two different shells.

flowchart TD
    Editor["Editor"] --> Studio["Prompt Studio\n(author + test)"]
    Studio --> Router["AI Router"]
    Router --> Anthropic
    Router --> Cerebras
    Router --> LMStudio["LM Studio"]
    Studio -.->|shares domain model| Workbench["Prompt Workbench\n(standalone web app)"]