Skip to content

Praneel7015/flowforge

BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

12 Commits
backend
backend
 
 
examples
examples
 
 
frontend
frontend
 
 
.gitignore
.gitignore
 
 
README.md
README.md
 
 
docker-compose.yml
docker-compose.yml
 
 
render.yaml
render.yaml
 
 

Repository files navigation

FlowForge

AI-powered low-code/no-code CI/CD automation platform

Build CI/CD pipelines visually with drag-and-drop, powered by AI. Platform-independent and supports multiple AI providers.

What is FlowForge?

FlowForge lets developers visually design CI/CD workflows using drag-and-drop nodes instead of manually writing configuration files. It uses AI to generate pipelines, analyze failures, auto-remediate broken pipelines, and answer questions about your CI/CD config in real-time chat.

Features

  • Visual CI/CD Workflow Builder - Drag-and-drop pipeline design with React Flow
  • AI Pipeline Generator - Describe a pipeline in natural language; AI writes the config and builds the visual graph
  • Multi-Platform Output - Generate configs for GitLab CI, GitHub Actions, Jenkins, or CircleCI
  • Multi-AI Provider Support - Use Featherless, Claude, Gemini, GPT-4, or local Ollama models
  • BYOM (Bring Your Own Model) - Users can provide their own API key/model/base URL from the UI
  • Pipeline Health Advisor - AI grades your pipeline (A-F) across speed, security, reliability, and best practices
  • Pipeline Chat (n8n) - Conversational assistant routed through an n8n webhook
  • Pipeline Migration - Upload a Jenkinsfile and convert it to any target platform
  • Platform-Independent - Not tied to any single CI/CD platform

Supported Providers

AI Providers

Provider Models
Claude (Anthropic) claude-opus-4-5, claude-sonnet-4-5
Gemini (Google) gemini-1.5-pro, gemini-1.5-flash
GPT-4 (OpenAI) gpt-4-turbo, gpt-4o
Featherless AI 30,000+ open models via OpenAI-compatible API
Ollama (Local) llama3, mistral, codellama

CI/CD Platforms

Platform Output File
GitLab CI .gitlab-ci.yml
GitHub Actions .github/workflows/ci.yml
Jenkins Jenkinsfile
CircleCI .circleci/config.yml

Architecture

Frontend (React + Vite + React Flow + TailwindCSS)
    │
    │ REST API
    ▼
Backend (Node.js + Express)
    │
    ├── AI Providers (Featherless, Claude, Gemini, OpenAI, Ollama)
    ├── CI/CD Generators (GitLab, GitHub, Jenkins, CircleCI)
    ├── Config API (/api/config/providers)
    └── Optional n8n chat webhook bridge (/api/advisor/chat)

Quick Start

Prerequisites

  • Node.js 18+
  • npm
  • At least one AI provider API key (or local Ollama)

1. Clone and install

git clone <repo-url> flowforge
cd flowforge

# Install backend
cd backend
cp .env.example .env
# Edit .env with your API keys
npm install

# Install frontend
cd ../frontend
npm install

2. Configure environment

Edit backend/.env:

# Choose your default providers
DEFAULT_AI_PROVIDER=featherless
DEFAULT_CICD_PLATFORM=github

# Add at least one AI provider key
ANTHROPIC_API_KEY=your_key_here
# Or use Gemini
GEMINI_API_KEY=your_key_here
# Or use OpenAI
OPENAI_API_KEY=your_key_here
# Or use Featherless (OpenAI-compatible)
FEATHERLESS_API_KEY=your_key_here
FEATHERLESS_BASE_URL=https://api.featherless.ai/v1
FEATHERLESS_MODEL=Qwen/Qwen2.5-Coder-1.5B-Instruct
# Or use local Ollama (no key needed)
OLLAMA_BASE_URL=http://localhost:11434

PORT=3001

n8n webhook chat integration (sponsor-ready)

FlowForge pipeline chat can be routed through n8n by setting these backend env vars:

N8N_CHAT_WEBHOOK_URL=https://your-n8n-domain/webhook-test/flowforge
N8N_WEBHOOK_AUTH_HEADER=X-FlowForge-Token
N8N_WEBHOOK_AUTH_TOKEN=your_secret_token
N8N_CHAT_TIMEOUT_MS=30000

FlowForge sends a JSON POST payload to this webhook with these key fields:

  • text / input: latest user message text
  • userId, username, chatId, conversationId: user-scoped session identifiers
  • currentYaml and yamlContext: full pipeline YAML context
  • messages: full chat message history for context
  • cicdPlatform, platformDisplayName, aiProvider, source, timestamp

For n8n workflows using the Webhook node output, map fields from body:

  • {{$json.body.text}}
  • {{$json.body.userId}}
  • {{$json.body.username}}
  • {{$json.body.chatId}}
  • {{$json.body.currentYaml}}

To keep chat memory personal per user, set your n8n memory session key to either {{$json.body.userId}} (all chats for a user) or {{$json.body.chatId}} (separate memory per conversation).

Expected n8n response can be either a plain string or JSON with one of these fields: reply, response, message, output, or text.

When N8N_CHAT_WEBHOOK_URL is not set, FlowForge falls back to its configured local AI provider.

3. Run

# Terminal 1 - Backend
cd backend
npm run dev

# Terminal 2 - Frontend
cd frontend
npm run dev

Open http://127.0.0.1:5174

Docker

docker-compose up

Usage

Visual Workflow Builder

  1. Drag pipeline nodes from the left sidebar onto the canvas
  2. Connect nodes by dragging from source handles to target handles
  3. Click a node to configure its settings (image, script, stage, etc.)
  4. Select your target CI/CD platform from the header dropdown
  5. Click Export to generate the configuration file

AI Pipeline Generator

  1. Click the AI Generator tab
  2. Select your AI provider and target CI/CD platform
  3. Describe your pipeline in natural language
  4. Click Generate Pipeline - the AI creates config and visual nodes
  5. The workflow loads automatically into the visual builder

Pipeline Migration

  1. Click the Migration tab
  2. Paste a Jenkinsfile or upload one
  3. Select your target platform (GitHub Actions, GitLab CI, etc.)
  4. Click Convert to migrate

Health Advisor

  1. Build any pipeline (drag nodes, use AI Generator, or import)
  2. Click Export on the canvas to capture the config
  3. Click the Health Advisor tab
  4. Click Analyse My Pipeline - AI scores it across 4 dimensions

Pipeline Chat

  1. Export a config first (so AI has context)
  2. Click the Pipeline Chat tab
  3. Ask anything: "Why might this be slow?", "Add caching", "Is this secure?"

Bring Your Own Model (BYOM)

  1. In the top bar, choose any AI provider (including Featherless)
  2. Enable Bring Your Own Model Key
  3. Paste your API key, and optionally override model/base URL
  4. Generate, migrate, score, and chat requests will use your credentials per request

Notes:

  • BYOM keys are not persisted in the backend
  • If BYOM is off, server-side env keys are used

Project Structure

flowforge/
├── frontend/
│   ├── src/
│   │   ├── components/
│   │   │   ├── Header.jsx            # Nav with provider selectors
│   │   │   ├── Sidebar.jsx           # Draggable node palette
│   │   │   ├── ProviderSelector.jsx  # AI/CI-CD provider dropdown
│   │   │   ├── BringYourOwnModelPanel.jsx # BYOM key/model/base URL
│   │   │   ├── PromptPanel.jsx       # Natural language generator
│   │   │   ├── JenkinsConverter.jsx  # Pipeline migration
│   │   │   ├── YamlPreview.jsx       # Config preview + download
│   │   │   ├── HealthAdvisor.jsx     # Pipeline health scoring
│   │   │   └── PipelineChat.jsx      # Conversational AI
│   │   ├── workflow/
│   │   │   ├── WorkflowEditor.jsx    # React Flow canvas
│   │   │   ├── NodeConfigPanel.jsx   # Node settings panel
│   │   │   └── nodes/
│   │   ├── App.jsx
│   │   └── main.jsx
├── backend/
│   ├── config/
│   │   └── index.js                  # Central config loader
│   ├── providers/
│   │   ├── ai/                       # AI provider implementations
│   │   │   ├── AIProvider.js
│   │   │   ├── FeatherlessProvider.js
│   │   │   ├── AnthropicProvider.js
│   │   │   ├── GeminiProvider.js
│   │   │   ├── OpenAIProvider.js
│   │   │   ├── OllamaProvider.js
│   │   │   └── index.js
│   │   └── cicd/                     # CI/CD generators
│   │       ├── CICDGenerator.js
│   │       ├── GitLabCIGenerator.js
│   │       ├── GitHubActionsGenerator.js
│   │       ├── JenkinsGenerator.js
│   │       ├── CircleCIGenerator.js
│   │       └── index.js
│   ├── controllers/
│   ├── routes/
│   ├── services/
│   │   ├── aiService.js              # AI operations + optional n8n chat bridge
│   │   ├── workflowService.js        # Nodes → CI/CD config
│   │   └── gitlabService.js          # GitLab API client
│   ├── utils/
│   │   └── aiOptions.js              # BYOM request option sanitizer
│   └── server.js
├── docker-compose.yml
└── README.md

API Endpoints

Method Endpoint Description
GET /api/config/providers List available AI/CI-CD providers
GET /api/config/providers/:type/:name/status Check provider status
POST /api/pipelines/generate AI pipeline generation from prompt (aiOptions supported)
POST /api/pipelines/export Export workflow nodes to config
POST /api/migration/jenkinsfile Convert Jenkinsfile to target platform (aiOptions supported)
POST /api/advisor/health Pipeline health score (aiOptions supported)
POST /api/advisor/remediate Auto-remediate broken config (aiOptions supported)
POST /api/advisor/chat Pipeline chat assistant (aiOptions supported, can route to n8n webhook)
GET /api/health Health check

Tech Stack

  • Frontend: React 18, Vite, React Flow, TailwindCSS, Axios
  • Backend: Node.js, Express, js-yaml
  • AI: Featherless AI, Anthropic Claude, Google Gemini, OpenAI GPT-4, Ollama
  • CI/CD: GitLab CI, GitHub Actions, Jenkins, CircleCI
  • Orchestration: n8n webhook integration for pipeline chat

License

MIT

About

flowforge-ten.vercel.app

Resources

Readme
Activity

Stars

1 star

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages