- web/: Local chat UI (Express + WS → Codex bridge) - openwebui/: Preset, pipelines, knowledge manifest - Dockerfile + docker-compose.yml - Updated README with 3 frontend options - CLI-agnostic: works with Codex, Claude Code, Kiro, Gemini
3.4 KiB
Reltio PM Factory
A browser-based AI workspace for Product Managers. No terminal required after initial setup.
Forge Console gives you a familiar chat interface backed by Codex, pre-loaded with Reltio's Transparent Factory standards, BMad Creative Suite, Jira/Aha!/Gainsight integrations, and epic-writing skills.
Quick Start
Option 1: Docker (recommended)
git clone <this-repo> pm-template
cd pm-template
cp env.example .env # fill in your API keys
docker compose up -d
Open http://localhost:3000.
Option 2: Local
git clone <this-repo> pm-template
cd pm-template
./start.sh
On first run, the script checks Node.js, creates .env, installs deps, and opens the browser.
Prerequisites (One-Time Setup)
Before the AI agent can connect to your tools, you need:
- Git & Bitbucket Access — A working SSH key configured with Bitbucket.
- Jira Authentication — Run
mcporter auth atlassianin a terminal once to complete the OAuth flow. - API Keys — Edit
.envwith your Aha! and Gainsight PX credentials.
Once authenticated, the Forge Console handles everything else.
What's Inside
| Component | Purpose |
|---|---|
web/ |
Forge Console — local chat UI + Express server |
skills/epics-standards/ |
Epic creation templates & Aha! workflows |
skills/factory-standards/ |
Transparent Factory tenet sync & compliance |
skills/bmad-suite/ |
BMad Creative Intelligence Suite (brainstorming, design thinking, storytelling) |
skills/gainsight-px/ |
Product analytics queries |
config/mcporter.json |
MCP server definitions (Jira, Aha!, Context7) |
For AI Agents
If you are an AI assistant reading this repo, see AGENTS.md for your operating instructions.
Frontend Options
This repo is CLI-agnostic. Use whichever interface fits your team:
Option A: Forge Console (built-in)
A lightweight local chat UI that wraps Codex CLI.
./start.sh # opens http://localhost:3000
Option B: Open WebUI (recommended)
Use the pre-built artifacts in openwebui/ to connect any Open WebUI instance.
Includes model presets, RAG knowledge manifests, and tool pipelines.
See openwebui/SETUP.md for details.
Option C: Any CLI Agent
Open the repo in Cursor, Claude Code, Kiro, Gemini CLI, or any agent that reads AGENTS.md.
Architecture
┌─────────────────────────────────────────┐
│ Frontend (pick one) │
│ • Forge Console (web/) │
│ • Open WebUI (openwebui/) │
│ • CLI Agent (Codex, Claude, Kiro...) │
└──────────────┬──────────────────────────┘
↕
┌──────────────┴──────────────────────────┐
│ LLM (any OpenAI-compatible provider) │
│ + System Prompt (AGENTS.md) │
│ + RAG Knowledge (skills/, standards/) │
│ + Tool Pipelines (mcporter, aha, etc.) │
└──────────────┬──────────────────────────┘
↕ MCP
Jira · Aha! · Gainsight PX