DEV Community: Shane

Build AI Chat Assistant App Using ReactJS & NodeJS Within 15 Minutes Using Automated AI Coding SaaS API From Your Browser

Shane — Mon, 28 Jul 2025 01:14:13 +0000

What I am about to demonstrate:
Building a custom AI Chat Assistant that is written in ReactJS and NodeJS without writing any code at all. It will all be handled by a new AI Automation Coding API Blaze Engineer. What will be built in the demonstration can easily be extended and/or modified.

Prerequisites - Which I have already completed:

Created an account at Blaze Engineer.
Logged into Blaze Engineer to access my Auth Token.
Generated a SSH Private/Public Key
Added SSH Public Key To Github Account
Added SSH Private Key To Blaze Engineer Account
Wrote two Blaze Engineer Coding Job Tasks.

The first Blaze Engineer Coding Job Task is huge because it defines the base design/architecture/features for the new project.

The second Blaze Engineer Coding Job Task is small because it is simply adding a new small feature to the just created new project.

Here is the brand new Github Repository before the initial Blaze Engineer Coding Job runs:

Left = Two Blaze Engineer Code Job Task Definitions in Notepad
Center = Blaze Engineer API Playground
Right = Brand New Github Repository

Before we can run the initial coding job on the brand new Github repository we must add the new repository to list of usable repositories in Blaze Engineer:

Now that we have added the new repository to Blaze Engineer. We can now run the initial coding job on the brand new repository. This is how you setup running a new job:

This is the response from running a new job:

Here is the Github files changes from the initial coding job:

Here is adding an additional new feature to the code base we just generated in the first coding job:

Here is the Github files changes from the add additional new feature coding job:

Here is the file changes in more detail:

Other ways of tapping into Blaze Engineer
HTTPS API
NodeJS Client Library NPM Package
Interactive CLI Tool NPM Package

Here is the Blaze Engineer API Automated Coding/DevOps Workflow Flow Chart:

We have it setup so when a task gets dropped in a certain list in our Task Board application it triggers a /jobs/run request, then waits for the request to complete, then run tests, then if the tests pass, it deploys, and then it will move the task in the Task Board to the Completed By Blaze Engineer list.

Conclusion

Writing the Coding Job Tasks is the most time consuming part of the 15 minutes. Blaze engineer does the rest quick. The initial code sync on a new repository and branch can take a few minutes but future coding job tasks on the same repository and branch will skip code syncing and jump straight to coding.

If the Coding Job Task is written without contradictions, missing critical details, and has enough details of the user's vision, it should with a high success rate do a good job of providing a good outcome.

Blaze Engineer just reached beta. So we have not got around to getting everything we can to make the costs as cheap as possible completed yet but we will. The cost right now are reasonable but we can make it even better!

We're looking for JavaScript/TypeScript engineers to test out the Beta of Blaze Engineer. If testing out Blaze Engineer sounds interesting to you then apply for beta access at Blaze Engineer and your application should be reviewed quickly. If granted access you will be emailed an beta key plus an invite to the Blaze Engineer Discord.

Happy vibe coding the easy way!

- The Blaze Engineer Team

Writing Good Blaze Engineer Tasks (Automate Coding Tasks)

Shane — Wed, 23 Jul 2025 21:58:33 +0000

Mastering Task Writing for Blaze Engineer

How to craft crystal‑clear coding jobs that unlock the full power of AI‑driven software engineering.

Quick Link: Explore the platform at blaze.engineer

Why Your Tasks Matter

Blaze Engineer turns natural‑language tasks into concrete code changes. The clearer your instructions, the faster the platform can ship working features. A well‑written task can save hours of back‑and‑forth, reduce hallucinations, and let the AI focus on what it does best: building software.

Every task you write is sent to the Blaze Engineer API as a Coding Job. The more context you provide, the tighter the feedback loop.

Blaze Engineer's Superpowers

Blaze Engineer excels at

Creating brand‑new projects from scratch.
Adding new features to an existing codebase.
Modifying existing features for enhancements or bug fixes.
Refactoring entire repositories for style, structure, or technology upgrades.

Give it a precise mission in any of these areas and the system will handle project scaffolding, file generation, and even heavy refactors on a live codebase.

Task Archetypes & What to Include

Below is a distilled version of the in‑app Writing Good Tasks guide.

1. New Project Tasks

When you need Blaze Engineer to spin up a green‑field repository, specify:

Stack - frontend, backend, or full‑stack.
Runtime details - Node version, JavaScript or TypeScript, ESM or CommonJS.
Frameworks/Libraries - React, Vue, Express, Fastify, etc.
External services - databases, caches, message buses, third‑party APIs.
Architecture & Design - desired folder layout, entry files, screen list, API routes.
Feature list - bullet‑point the MVP scope.

Example: "Create a full‑stack video‑sharing platform… React SPA (no JSX), Express backend, MongoDB + AWS S3…"

2. Editing Existing Projects

Adding a New Feature
- Provide a short summary plus whether it touches UI, API, or both. Link to relevant files or folders when possible.
Extending an Existing Feature
- Describe what you're adding and which feature it augments. Highlight edge cases.
Project‑Wide New Capability
- Outline the cross‑cutting concern (e.g., internationalization, logging, dark mode) and enumerate the file types Blaze Engineer should modify.
Refactors
- State the refactor's goal - naming, formatting, performance, or tech migration - and scope (entire repo or specific modules).

Five Golden Rules for Better Tasks

Lead with a one‑sentence summary. Give Blaze Engineer the TL;DR up front.
List environment variables and external service credentials it will need.
Describe folder structure or point to existing directories for context.
Use bullet points for features, endpoints, or UI screens - AI parses lists quickly.
Clarify output expectations. If you want every changed file returned, say so.

Example Tasks From the Official Guide

The HTML guide ships with fully‑fleshed samples. Here are two that illustrate what "good" looks like.

Example 1 - Kick‑Off a New Project

We need to create the base of a new project. The new project is a Simple Video Sharing Platform. The Frontend will be a ReactJS Application not using JSX. The Backend will be an ExpressJS server that serves the ReactJS Application and the endpoints the Frontend Application will make requests too.

The Backend will upload the video file to an AWS S3 Bucket then insert a MongoDB record in a Videos Collection containing the title, description, uploadedTime, and location using the mongodb NPM Package and using the simpleVideoSharingPlatform MongoDB Database. The location will be the URL to the Video File in the AWS S3 Bucket. Use dotenv NPM Package to load environment variables. Create and generate the src/index.js file that will configure, setup, and run ExpressJS server. Create and generate every other file needed for the basic functionality for each Backend feature. The Backend will use IMPORT instead of require to import files.

The Frontend will have an Upload Video Screen and a View Uploaded Videos Screen. The two screens should be contained inside a single file ReactJS html file named core.html that can be served as a SPA that will contain all the HTML/CSS/JavaScript for everything inside the SPA.

Here is the folder structure I want the project to start with:

/src
 /classes
 /functions
 /endpoints
 /ui
  core.html
 index.js

There are no user accounts or authorization. Any visitor can upload a video or view the uploaded videos. The data attached to an uploaded video will be the title, description, uploadedTime, and the location. The visitor provides the title and description. The system provides the current time for the uploadedTime and the AWS S3 Bucket URL for the Video File location.

Example 1 Output

Example 2 - Editing an Existing Project

Add a delete‑button feature to uploaded videos. The button should appear-and work-only for visitors whose IP matches the IP of the uploader.

Example 2 Output

Notice how each example:

Starts with a plain‑language summary.
Specifies the what (feature), where (code areas), and any constraints (IP match).
Provides folder layouts or environment hints when relevant.

Use these as templates and adapt them to your own domain.

Still in Beta - Grab Your Free Credits

The Blaze Engineer public beta is live and currently accepting beta‑key requests. Approved testers receive $10 in free AI credits to explore automated project creation, feature work, and repo‑scale refactors. Spaces are limited - request your key while the program is open!

Final Thoughts

Task writing is a skill, and with Blaze Engineer the payoff is exponential. By sharpening your prompts, you let the platform do what it does best - generate, modify, and refactor code at machine speed so you can ship faster.
Ready to give it a try? Draft a task, send it to the Blaze Engineer API, and watch your next feature land in minutes.