# Browser Research Clipper with CSV/JSON Export

Project type: Personal project — browser extension prototype

Role: AI-assisted project development and workflow iteration

Suggested skills: JavaScript, Google Chrome Extension, Data Extraction, JSON, CSV. Use matching tags offered by the platform.

## Short description

A Manifest V3 extension prototype for saving research from the active page into a local library. The source includes page-title and excerpt capture, editable review fields, keyword search, and CSV/JSON export. Core data and export logic was checked with two synthetic records. Browser APIs were mocked; installation and live active-tab capture were not verified in this review.

## Client problem

Manual web research often leaves useful excerpts scattered across tabs and notes. A small local capture tool can keep the original URL, excerpt and human review notes together.

## Implementation

The extension declares activeTab, scripting and storage permissions. A user-triggered capture reads the current page title, selected text or an excerpt, and fills editable fields. Records are saved in chrome.storage.local. The library includes text search, copying, removal and CSV/JSON exports.

## Evidence

- Eight focused checks passed for normalization, fallback values, CSV quoting, empty-state handling and serialization.
- JSON export retained both synthetic records.
- CSV export retained commas and embedded quotation marks with a UTF-8 BOM.
- Sample exports came from the original export function with the download boundary intercepted.

See `../samples/clipper-export.json` and `../samples/clipper-export.csv`. These are fictional records, not live page captures.

## Appropriate small engagement

Add an agreed form field, adjust an export schema, or implement a narrow user-triggered page-capture workflow. Validate the target website and browser first.

## Scope and limitations

The existing interface and CSV headers are Chinese. This is not an unattended crawler, cloud CRM or published browser-store product. Browser-restricted pages may not be readable. Actual extension permissions, persistence and active-tab capture still need end-to-end browser verification. CSV formula-like values are not neutralized by the current export function; handling untrusted spreadsheet input needs an additional change before broader use.

Development used Codex assistance. This is a personal project, not a paid-client case study.