AssembleAI

No-Code Drag-and-Drop AI Workflow Studio
Premium dark SaaS MVP for building, testing, and reviewing auditable AI workflows.

Overview

AssembleAI is a product-style AI workflow platform for visually designing and testing document and text evaluation pipelines.

The current build includes:

• Premium landing page and auth UI
• Dashboard with workflows, templates, and recent operational activity
• Workflow Builder with React Flow canvas, node library, inspector, execution preview, and builder assistant
• Runtime Workbench with input composer, stage insights, execution preview, evaluation result surfaces, and contextual assistant
• Runs, reviews, and templates product areas
• Server-side OpenAI evaluation and grounded follow-up chat with graceful fallback when no API key is configured

Core Workflow Example

The flagship demo flow is:

File Input -> OCR -> NLP -> Rules -> Rubric Scoring -> GPT -> Output

This is modeled around a Paper Evaluation AI use case for assessing answer sheets with rubric-based scoring, grounded feedback, and human review readiness.

Tech Stack

• Next.js App Router
• TypeScript
• Tailwind CSS
• Framer Motion
• React Flow
• Zustand
• Lucide Icons
• Zod
• OpenAI Node SDK

Product Areas

Builder

• Drag-and-drop workflow studio
• Resizable node library and inspector
• Demo workflow preloaded
• Builder test-run input composer
• Full-width processing stages, execution preview, and builder assistant

Workbench

• Workflow-aware input composer
• File upload plus instructions and advanced context
• Full-width runtime pipeline stages
• Execution preview with tabs for trace, output preview, evidence, warnings, and logs
• Full-width contextual assistant grounded to the active run

Review and Audit

• Runs listing and run detail views
• Review queue and review detail flows
• Template gallery and versioning surfaces

Project Structure

src/
  app/
    api/
    builder/
    dashboard/
    login/
    reviews/
    runs/
    templates/
    workbench/
  components/
    builder/
    layout/
    reviews/
    runs/
    shared/
    templates/
    ui/
    workbench/
  data/
  lib/
    ai/
    review/
    validation/
    workflow/
  store/
  types/

Running Locally

1. Install dependencies

npm install

2. Configure environment variables

Create a local .env file from the example:

cp .env.example .env

Then configure:

OPENAI_API_KEY=your_key_here
OPENAI_MODEL=gpt-4o-mini

Notes:

• OPENAI_API_KEY is optional for basic local demo usage
• if the key is missing, the app falls back to local simulated evaluation behavior
• OpenAI is used only from server-side routes and services

3. Start the dev server

npm run dev

Open:

http://localhost:3000

4. Quality checks

npm run lint
npm run build

Environment and Security

• .env is intentionally ignored and should never be committed
• .env.example is safe to commit and documents required variables
• API keys are read only on the server
• Client components never receive the OpenAI secret

API Endpoints

POST /api/evaluate

Server-side workflow evaluation endpoint.

Accepts:

• workflow metadata
• extracted input text
• rubric
• rules
• optional expected/reference answer

Returns structured evaluation output including:

• overall score
• rubric breakdown
• strengths and weaknesses
• missing concepts
• improvement suggestions
• confidence
• review routing metadata

POST /api/chat

Server-side grounded follow-up chat endpoint.

Accepts:

• active run context
• user question
• prior chat messages

Returns:

• grounded answer
• citation paths / run references
• suggested next questions

Current Mocked vs Real

Real

• App shell and product UI
• Server-side OpenAI integration
• Typed evaluation and chat contracts
• Workflow builder and runtime UX
• Review and run surfaces

Mocked / Simplified

• OCR depth
• persistent database storage
• auth backend
• deployment pipeline
• multi-user collaboration

GitHub Setup

This project is intended to be pushed to:

https://github.com/Mithran-MV/AssembleAI.git

If you want to initialize and push manually:

git init
git branch -M main
git remote add origin https://github.com/Mithran-MV/AssembleAI.git
git add .
git commit -m "Initial commit"
git push -u origin main

Recommended Next Steps

• Add persistent workflow and run storage
• Add real document parsing and OCR adapters
• Add authenticated reviewer identities and protected routes
• Add workflow publishing and deployment flows
• Add export/download packages for run evidence and review reports