Skip to content

About

Delivery planning workspace with KPI calculators, risk tracking, persistent drafts and streaming AI chat.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

10 Commits

Folders and files

Repository files navigation

Delivery Workbench

A delivery-planning workspace for KPI calculations, risk assessment, sprint capacity, estimates, and AI-assisted analysis.

Delivery Workbench brings a set of related planning tools into one keyboard-accessible interface. It combines a React client with an Express server, local SQLite persistence, and optional streaming Gemini chat. The project demonstrates a compact full-stack application that can run as a single service.

Delivery Workbench showing its KPI calculator workspace

Actual local application with seeded example data. Previously named AKB1 Command Center.

Status: portfolio prototype. The previous hosted demo is unavailable; the local walkthrough below is the supported starting point.

Run locally · Architecture · Code guide

Capabilities

Area Included tools
Delivery planning Sprint capacity, estimation, pricing, decision comparison
Reporting KPI calculators, risk register, status-report drafts
AI assistance Streaming chat with saved sessions and a delivery-oriented system prompt
Prompt tooling Prompt templates, a structured prompt builder, and model guidance
Workspace Command palette, responsive navigation, persisted tool drafts

Calculators and reports operate on entered values and seeded examples. They are not automatic integrations with a company's delivery systems. Without a Gemini key, chat uses a demonstration response.

Architecture

flowchart LR
    User["Browser"] --> UI["React, Vite and TanStack Query"]
    UI -->|HTTP and SSE| API["Express application"]
    API --> Store["Storage layer: node:sqlite"]
    Store --> DB[("SQLite file")]
    API -->|configured chat requests| Gemini["Gemini API"]
    Gemini -->|streamed response| API
    API -->|no API key| Demo["Demo response"]
    API --> Static["Built client assets"]
Loading

Persistence uses Node's built-in node:sqlite directly. Drizzle-related files remain in the repository, but the running storage layer is server/storage.ts. Express serves the API and the built frontend in production; Vite is attached during development.

Run locally

Requires Node.js 25.6.0 or newer, as specified by the package, and npm.

git clone https://github.com/anudeepadi/delivery-workbench.git
cd delivery-workbench
npm ci
cp .env.example .env
npm run dev

Open localhost:5000. SQLite tables and seed data are initialized by the storage layer. No external database service is required.

Variable Default / purpose
PORT 5000
HOST Loopback in development; all interfaces in production
SQLITE_DB_PATH data/akb1.sqlite
GEMINI_API_KEY Optional; enables provider-backed chat
GEMINI_MODEL gemini-2.5-flash, subject to provider availability

The server loads .env at startup. Chat requests send conversation context to Gemini when configured; saved history remains in the application's database.

Five-minute walkthrough

  1. Select Open Workbench and choose KPI Engine.
  2. Change a calculator input and inspect the formula and calculated result.
  3. Open Risk Matrix or Sprint Planner to explore the sample planning tools.
  4. Edit a tool draft, reload the page and confirm it is restored from SQLite.
  5. Open chat without setting a key to see the clearly labeled demo response.

There is no live connection to a company delivery system. Seeded figures and benchmarks are illustrative and should be reviewed for your context.

Build and check

npm run check
npm run build
npm start

Type checking, the production build, health endpoint, local draft persistence and no-key chat were checked on 22 September 2026 with Node.js 26.9.0. No Gemini requests were made.

GET /api/health is the health endpoint. There is no test command declared in the package; type checking and building are the available baseline checks.

For a container deployment, use the included Dockerfile and persist the SQLite directory on a volume. A disposable container filesystem will not preserve chat history or drafts across replacements.

Code guide

Path Responsibility
client/src/components/tabs/ Planning and reporting modules
client/src/hooks/use-tool-draft.ts Draft persistence from the UI
server/routes.ts Bootstrap, sessions, drafts, and streaming chat
server/storage.ts SQLite schema, seeds, and data access
shared/contracts.ts Client/server domain contracts
script/build.ts Frontend and server build

Scope and contributions

This is a portfolio application with a shared workspace, not a tenant-isolated enterprise product. The current routes do not provide per-user authentication and authorization. Keep personal experimentation local; hosted multi-user use needs an access-control design.

Useful contributions include calculator examples, input validation, accessibility improvements, and tests for persistence and streaming. Describe the affected module and include steps to reproduce any behavior change.

License

package.json declares MIT, but a standalone license file is not currently tracked. Clarify repository-wide licensing before redistributing a release.

About

Delivery planning workspace with KPI calculators, risk tracking, persistent drafts and streaming AI chat.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages