← Portfolio index
003dev

Portfolio Workbench

Portfolio Workbench is a local-first web application for documenting product ideas, experiments, prototypes, and shipped work using Markdown and YAML files stored in Git. It renders a searchable, filterable portfolio site with Astro, validates project metadata at build time, and includes a development-only editor for creating projects and timeline events without a database or external services. All data lives in version-controlled files, making the portfolio durable, inspectable, and portable for future deployment.

Why it was built

Built to keep project documentation and evidence tied to plain files under Git rather than a hosted service, so records stay durable, diffable, and portable, per the README's stated rationale ("Markdown and YAML in Git are the canonical data"). The local-only editor, loopback-request checks, and static build output show a deliberate choice to avoid a database or auth backend for the read path. The AI-fill and device-unlock features suggest the goal was to lower the effort of writing up new projects while still letting some entries stay restricted to the author via passkey rather than a login system.

Evidence from the work

Images, documents, and recordings stored alongside this project.

Delete this artifact?

will be removed from the project folder. This cannot be undone.

GitHub connection

Not connected.

Paste a fine-grained personal access token with Contents: read and write onaviman1258/project-workbench (add workbench-private for private projects). It is stored only in this browser.Create one on GitHub ↗

AI drafts (optional): paste anAnthropic ↗(sk-ant-…) orOpenAI ↗(sk-…) API key — the provider is detected automatically.

OpenAI keys additionally need the repo's proxy/openai-worker.js deployed on Cloudflare Workers (OpenAI blocks direct browser calls) — paste the worker URL here. Anthropic keys need no proxy.

Avishek Chandra

Product & engineering

achandra1258@gmail.com720-320-2111

Certifications

AWS Certified Cloud PractitionerAmazon Web ServicesDatadog Certified: Datadog FundamentalsDatadog

Scan to open