Skip to content

The only "Idea-to-Deployment" engine that builds with compliance baked in.

License

Notifications You must be signed in to change notification settings

cornerbrooke/ibm-trustbuild

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

2 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

IBM TrustBuild

The only "Idea-to-Deployment" engine that builds with compliance baked in.

IBM watsonx Langflow IBM Granite IBM Cloud


πŸ† IBM Dev Day: AI Demystified β€” Hackathon Submission

Theme: From Idea to Deployment

Team: cornerbrooke


πŸ“Œ The Problem

The biggest bottleneck in the enterprise isn't writing code β€” it's approval. Developers want to move fast, but security and compliance teams act as the "brakes" because AI-generated code can be risky or misaligned with cloud best practices. The gap between idea and deployed, compliant product costs teams weeks.

πŸš€ The Solution

IBM TrustBuild bridges this gap by combining a generative Architect Agent with a real-time Governance Guardrail into a single, end-to-end pipeline.

For Developers For the Enterprise
Automates architecture scaffolding and boilerplate using IBM Granite Ensures every generated line of code passes a Pre-flight Governance Audit
Generates Dockerfiles, Terraform, and IBM SDK integrations in seconds Enforces encryption, VPC isolation, and policy compliance automatically
One prompt β†’ deployment-ready kit Innovation and regulation live in the same workflow

🧠 Architecture: The TrustBuild Pipeline

β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚                        USER PROMPT INPUT                             β”‚
β”‚        "Build a customer portal for sensitive health data"          β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                            β”‚
                            β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚  STAGE 1: INTENT EXTRACTION          [granite-3-8b-instruct]        β”‚
β”‚  ─────────────────────────────────────────────────────────────────  β”‚
β”‚  β€’ Parses natural language into structured requirements              β”‚
β”‚  β€’ Identifies: stack needs, data sensitivity, scale requirements    β”‚
β”‚  β€’ Output β†’ requirements.json                                       β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                            β”‚
                            β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚  STAGE 2: ARCHITECT NODE (The Builder)                              β”‚
β”‚  ─────────────────────────────────────────────────────────────────  β”‚
β”‚  β€’ Maps requirements β†’ IBM Cloud services                           β”‚
β”‚  β€’ Selects: Code Engine, Cloudant, watsonx.ai, Container Registry   β”‚
β”‚  β€’ Output β†’ architecture_manifest.json                              β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                            β”‚
                            β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚  STAGE 3: GOVERNANCE GUARDRAIL (The Judge)   ⚑ INNOVATION LAYER   β”‚
β”‚  ─────────────────────────────────────────────────────────────────  β”‚
β”‚  β€’ Auditor Agent scans manifest against Policy Knowledge Base       β”‚
β”‚  β€’ Checks: encryption, VPC isolation, HIPAA, public API usage       β”‚
β”‚  β€’ ❌ FAIL β†’ Auto-corrects architecture before code generation      β”‚
β”‚  β€’ βœ… PASS β†’ Approves for code synthesis                            β”‚
β”‚  β€’ Output β†’ governance_report.json                                  β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                            β”‚
                            β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚  STAGE 4: SECURE CODE SYNTHESIS          [granite-20b-code]         β”‚
β”‚  ─────────────────────────────────────────────────────────────────  β”‚
β”‚  β€’ Generates Dockerfile, Terraform, app boilerplate                 β”‚
β”‚  β€’ Pre-configured with IBM SDKs and security policies               β”‚
β”‚  β€’ Output β†’ Deployment Kit (downloadable)                           β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

πŸ› οΈ Technology Stack

Component Technology Role
Reasoning Engine watsonx.ai (Granite 3.0) Powers all agent reasoning
Intent Parsing Granite-3-8b-instruct Decomposes prompts into requirements
Code Generation Granite-20b-code Generates Dockerfiles, Terraform, boilerplate
Orchestration Langflow Connects the 4-stage multi-agent pipeline
Governance watsonx.governance (simulated) Policy enforcement and audit logic
Deployment Target IBM Cloud Code Engine, Cloudant, Container Registry
Frontend React + TypeScript Dashboard UI with real-time pipeline visualization
Backend Python (FastAPI) API layer connecting frontend to watsonx

πŸ“‚ Repository Structure

ibm-trustbuild/
β”œβ”€β”€ README.md                   # This file
β”œβ”€β”€ docker-compose.yml          # Full-stack local deployment
β”œβ”€β”€ .env.example                # Environment variable template
β”‚
β”œβ”€β”€ frontend/                   # React + TypeScript dashboard
β”‚   β”œβ”€β”€ package.json
β”‚   β”œβ”€β”€ tsconfig.json
β”‚   β”œβ”€β”€ public/
β”‚   β”‚   └── index.html
β”‚   └── src/
β”‚       β”œβ”€β”€ index.tsx
β”‚       β”œβ”€β”€ App.tsx             # Root app with routing
β”‚       β”œβ”€β”€ components/
β”‚       β”‚   β”œβ”€β”€ Header.tsx      # Top nav with branding
β”‚       β”‚   β”œβ”€β”€ Pipeline.tsx    # 4-stage pipeline visualization
β”‚       β”‚   β”œβ”€β”€ PromptInput.tsx # User prompt entry
β”‚       β”‚   β”œβ”€β”€ GovernancePanel.tsx  # Guardrail status display
β”‚       β”‚   └── OutputKit.tsx   # Final deployment kit display
β”‚       └── styles/
β”‚           └── globals.css     # IBM Carbon-inspired theming
β”‚
β”œβ”€β”€ backend/                    # Python FastAPI API server
β”‚   β”œβ”€β”€ requirements.txt
β”‚   β”œβ”€β”€ main.py                 # FastAPI app entry point
β”‚   β”œβ”€β”€ routes/
β”‚   β”‚   └── pipeline.py         # /api/run-pipeline endpoint
β”‚   β”œβ”€β”€ agents/
β”‚   β”‚   β”œβ”€β”€ intent_agent.py     # Stage 1: Intent Extraction
β”‚   β”‚   β”œβ”€β”€ architect_agent.py  # Stage 2: Architecture Mapping
β”‚   β”‚   β”œβ”€β”€ governance_agent.py # Stage 3: Governance Guardrail
β”‚   β”‚   └── codegen_agent.py    # Stage 4: Secure Code Synthesis
β”‚   β”œβ”€β”€ services/
β”‚   β”‚   β”œβ”€β”€ watsonx_client.py   # watsonx.ai API wrapper
β”‚   β”‚   └── policy_kb.py        # Governance Policy Knowledge Base
β”‚   └── models/
β”‚       └── schemas.py          # Pydantic request/response models
β”‚
β”œβ”€β”€ langflow/                   # Langflow pipeline export
β”‚   └── trustbuild_pipeline.json # Importable Langflow flow definition
β”‚
β”œβ”€β”€ scripts/                    # Utility and deployment scripts
β”‚   β”œβ”€β”€ setup.sh                # One-command local environment setup
β”‚   └── deploy_ibm_cloud.sh     # IBM Cloud deployment script
β”‚
β”œβ”€β”€ docker/                     # Docker configurations
β”‚   β”œβ”€β”€ Dockerfile.frontend     # Frontend container
β”‚   └── Dockerfile.backend      # Backend container
β”‚
└── docs/                       # Supporting documentation
    └── DEMO_SCRIPT.md          # 2-minute demo video script

⚑ Quick Start

Prerequisites

  • Node.js 18+ and npm
  • Python 3.10+
  • Docker & Docker Compose (optional, for full-stack)
  • IBM Cloud account with watsonx.ai access
  • watsonx.ai API Key and Project ID

1. Clone & Configure

git clone https://github.com/cornerbrooke/ibm-trustbuild.git
cd ibm-trustbuild

# Copy and edit the environment file
cp .env.example .env
# Edit .env with your IBM watsonx credentials

2. Run with Docker Compose (Recommended)

docker-compose up --build

3. Run Locally (Without Docker)

# Terminal 1: Frontend
cd frontend
npm install
npm run dev

# Terminal 2: Backend
cd backend
pip install -r requirements.txt
uvicorn main:app --reload

🎬 Demo

See docs/DEMO_SCRIPT.md for the full 2-minute recorded demo script.

Live Demo Flow:

  1. Open the TrustBuild dashboard
  2. Enter: "Build a customer portal that uses AI to analyze sensitive health data"
  3. Watch the 4-stage pipeline execute in real time
  4. Observe the Governance Guardrail flag and auto-correct a HIPAA compliance issue
  5. Download the final Deployment Kit

πŸ“‹ Judging Criteria Alignment

Criteria How TrustBuild Scores
Completeness & Feasibility (5pts) Full 4-stage pipeline with working frontend, backend, and Langflow export. Real watsonx.ai integration via API.
Creativity & Innovation (5pts) The Governance Guardrail with auto-correction is the differentiator. No other "code generator" enforces compliance before synthesis.
Design & Usability (5pts) IBM Carbon-inspired UI with real-time pipeline status, animated stage transitions, and a one-click deployment kit download.
Effectiveness & Efficiency (5pts) Directly addresses the hackathon theme. Reduces idea-to-deployment from days to 60 seconds. Scales to any enterprise policy rulebook.

πŸ“„ License

This project was built for the IBM Dev Day: AI Demystified Hackathon (January 2026).


Built with watsonx. Governed by design.

About

The only "Idea-to-Deployment" engine that builds with compliance baked in.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published