JobWeave

Extension UI
More UI

Inspiration

Job seekers waste hours tailoring resumes and drafting cover letters.
Recruiters and ATS reward precise phrasing and mapped experience.
We wanted to remove busywork and return control to applicants while protecting their data.

What it does

Our extension provides a solution for job seekers. When a user opens a job posting and launches our extension, they can extract the job description and instantly receive tailored resume points and a matching cover letter. Additionally, users can upload their existing resume to get a rewritten, job-specific version optimized for maximum selection probability. What once took hours can now be done in seconds with our extension.

How we built it

Browser extension as the UI layer. Content scripts parse job pages and extract structured fields.
Front-end uses a lightweight React interface for real-time preview.
Chrome's APIs in the backend handles communication with the on-device Gemini Nano model.
User data stored locally by default. Cloud LLM for opt-in features and high performance resume improvements (only if users consents).

Challenges we ran into

Platform & API constraints

Chrome built‑in AI APIs require explicit outputLanguage, user‑gesture for model create when model is "downloadable/downloading", and aren’t available in service workers. These rules caused several runtime DOMException/user‑gesture bugs and required careful gating of create() calls.

Origin trials & availability differences

Writer/Rewriter/Proofreader are behind origin trials/flags on many dev machines; had to build graceful degradation so Summarizer‑only paths still work and surface clear guidance for enabling flags.

High System Requirements and Client Run

System requirements are high for Built-in AI and only one of our computer could run it. Then too, performance could vary depending upon available storage, memory and VRAM.

Accomplishments that we're proud of

Working prototype that extracts job descriptions and produces tailored resume bullets and a cover letter in seconds.
Demonstrated measurable time savings by replacing multi-hour manual edits with instant rewrites.
Clear, testable prompt templates that produce recruiter-friendly language.
Privacy-first architecture that gives users control over sensitive fields.

What we learned

Extraction reliability beats perfect parsing. Get usable structured data quickly.
Constrained LLM outputs are more trustworthy than unconstrained freeform generations.
Users prefer transparent edits. Show changes to the users.
Security and clear UX for consent are essential for adoption.

What's next for JobWeave

In-extension resume studio

Users will edit, version, and A/B test resumes inside JobWeave.
AI suggestions appear inline. Users accept or reject with one click.
Version history and export to ATS-ready formats turn casual users into repeat, paying customers.

One-click apply and workflow automation

Apply from the extension with a single validated submission.
Automated form-fill, resume selection, and cover-letter injection collapse the apply funnel from minutes to seconds.
That workflow drives frequency and creates monetizable usage events.

Personalized, AI-driven job feed and intent signals

Behavioral signals power a personalized feed. If a user applies mostly to SDE roles, we surface SDE openings first.
Signal aggregation builds a high-quality intent graph. That graph is a recurring-revenue asset for premium matching, alerts, and recruiter integrations.

Subscription tiers

Free tier for basic rewrites and one-off uses.
Premium tiers for prioritized rewrites, multi-resume management and most importantly providing access to paid cloud hosted LLM that doesn't use chat data for training.
Enterprise plans offer ATS integrations, bulk rewriting, and white-label deployments.