AI Initiative · Open source

Chrysalis — an AI chief of staff for your career move.

Job search is not one task — it's a system. Chrysalis is a personal AI operating system that keeps your pipeline, company research, interview prep, market signals, debriefs, and daily planning in a repo-backed knowledge base. Local-first. Claude Code-first. Your data stays in your own repo.

Chrysalis dashboard — pipeline, calendar, prep readiness, and industry pulse in one view

The Chrysalis dashboard — generated locally from your repo on every brief.

The mental model

A calendar tells you when you're free. A chief of staff tells you what matters most in that free time.

That's the bar Chrysalis is aiming for: using context from your inbox, calendar, recruiter conversations, and pipeline to think strategically, not just tactically. Given who you are, where you are in each process, and what changed since yesterday — what should you focus on today?

What it does

The MVP, today.

Onboard from your resume

Fork the repo, run Claude Code, and say Onboard me. Chrysalis builds your profile, wires up email + calendar, and sweeps the last 14 days to auto-extract every company in flight.

Daily Brief

Say Brief each morning. Chrysalis reads your pipeline, scans industry news, flags stale cards, and tells you the single most important next action.

Round-specific interview prep

Prep for [company] researches that company's actual interview process (not a template), maps your stories to what the round tests, and builds a dated study plan around your calendar.

Debrief that compounds

After every round, capture what landed and what to fix in under 2 hours. Each company card gets richer; each next round's prep gets sharper.

Email + calendar aware

Tiered MCP routing — first-party providers, Composio, community MCPs, or paste mode. Auto-creates prep blocks before interviews and debrief blocks right after.

Self-contained dashboard

Every brief regenerates dashboard.html — kanban pipeline, week density, industry pulse, prep readiness. No server, no login, no extra app.

The commands

Plain English, every day.

You don't learn an API. You talk to it like a chief of staff.

Brief

Morning pipeline health + industry pulse + today's next action

Research [company]

Profile-aware web research → card written and committed

Prep for [company] interview

Process research → round prep → calendar-aware study plan

Debrief [company]

Post-round capture; next round details + next prep offered

When can I schedule [company]?

Scheduling rec based on prep load and downstream commitments

Am I over-scheduled?

Pipeline density and prep-coverage report

Inside Chrysalis

A look at the system.

Company card — interviewing

Company card — interviewing

Each company gets a living card: status, prep readiness, next action, business overview, and recent news from research sweeps.

Profile-aware research

Profile-aware research

Research isn't a template dump — it's mapped to your background. Geography, product × user segments, strategic priorities.

Background sweep routine

Background sweep routine

Email + calendar signals swept every 4 hours. New interviews, recruiter follow-ups, and confirmations land in the system automatically.

Setup is a sentence

Setup is a sentence

Just say "Setup the automatic sweep for every 4 hours." Chrysalis wires the scheduled task itself.

Open source on GitHub

Open source on GitHub

Public repo at github.com/notbotanand/Chrysalis — MIT licensed, fork it and make it yours.

Why open source

This workflow should be inspectable, forkable, and adaptable.

Every career move is different. Chrysalis isn't a closed black box — it's a framework you can read, modify, and make your own. The public MVP is Claude Code-first; next, I'm making it portable to ChatGPT so your career context belongs to you, not to a single assistant.

Try it, fork it, ship a better version.