Assist or Delegate? Choosing an AI Development Tool
Compare assistive coding tools with agentic development environments to see where human control, automation, verification, and evaluation belong in the workflow.
Loading post…
JetbrainsAliaksei KartynnikTutorial
Your diff has a demo now with Junie /demo
Summary
Junie /demo adds a mode to Junie CLI that builds and launches an application, drives its UI, and records a requested check. It produces an HTML report, screenshots, and video, while leaving reviewers able to inspect steps, outcomes, and evidence. In the example, a prompt selects two issues, marks them Done, checks counter changes, reloads the page, and verifies persistence. Initial setup analyzes the repository and generates configuration after a confirmed build and launch plan; Docker-based environments can use custom templates for complex repositories. The mode also supports GitHub Actions workflows for pull-request demonstrations and release smoke tests, but an explicit PASS verdict is required for a scenario to pass.
Context
The post presents /demo as a way to delegate repeatable UI verification while keeping recordings and reports available for review. It covers local setup, pull-request artifacts, release smoke tests, Docker-based environments, supported Computer Use models, and run costs.
Approach / What changed
Start Docker, ask Junie to set up /demo, review and confirm its proposed build and launch plan, then inspect the generated configuration. Run /demo against selected changes with a prompt that names the actions, expected results, and persistence checks. For CI, use –demo or the supplied GitHub Actions workflows, configure repository secrets and artifacts, and require an explicit PASS verdict for successful scenarios.
Takeaways
A specific prompt should state the UI actions, expected state changes, and conditions that must survive a reload; vague requests leave more room for interpretation.
The demo environment uses a Debian Bookworm Docker container with Chromium, Node.js, Xvfb, a window manager, xdotool, screenshot tools, and ffmpeg. Projects add dependencies through a Dockerfile and describe launch steps in .junie/demo.md.
A completed agent process is not itself a passing check in CI. The example accepts only an explicit PASS verdict; FAIL, PARTIAL, or missing or invalid verdicts fail the result while other scenarios can still upload evidence.
FAQ
What does Junie /demo produce after exercising an application?
It produces an HTML report, screenshots, and a video. The report includes the request, result, video, screenshots, executed steps, and checks marked passed, failed, or incomplete.
How does /demo work with complex repositories or monorepos?
A repository can define several VM templates with Dockerfiles under .junie/vms/. The .junie/demo.md file specifies which template to use, required services, and startup commands so Junie can select the appropriate environment.
What makes a CI demo scenario pass?
The prompt must ask Junie to write an explicit verdict. Only PASS passes the result check; FAIL, PARTIAL, and missing or invalid results fail it, although other scenarios may still complete and upload their evidence.
Source and freshness
- Source
- blog.jetbrains.com
- Published
- Added to Yomu
Summaries are generated from the original by a language model. Check the original before relying on it.
See the idea
Visual explainers
Keep reading
Related reading
The Good Stuff
Paid picks we use ourselves. Links carry a referral code — a subscription through one earns us a commission at no extra cost to you.
The best engineering reads, once a week
One useful email. Unsubscribe anytime.