Score postings against your compass. Discover companies. Teach the AI to know you. Own every word.
github.com/kandizzy/possible-futures →
Next.js · SQLite · Claude API · Local-first
mkdir career && cd career
git clone https://github.com/kandizzy/possible-futures.git
cd possible-futures && npm install
echo "ANTHROPIC_API_KEY=your-key-here" > .env.local
npm run dev
Open http://localhost:3000 and the intake begins.
Clone into a parent folder (career/, resume/, etc.) — the app writes your generated files there. New to the terminal, Node, or nvm? That's fine. The README walks through installing each piece step by step, and getting your Anthropic API key.
30 minutes, your last few roles in front of you, and the promise that you're writing this yourself.
Name, education, and the through-line that connects your work.
Two or more roles. Title, company, dates, summary, proudest moment. Messy notes welcome.
Dream role, role tiers, target companies, signal words, compensation floor.
Red flags, banned words, voice calibration. Click phrases that sound like you, or paste your own.
Projects, recognition, and resume versions. Add or remove versions as needed.
Different roles call for different resumes. Name them, describe their emphasis.
Preview your Book, Compass, and Playbook in rendered markdown or source view.
Your scoring rubric. Signal words, red flags, role tiers. Every posting is evaluated against this.
Three markdown files saved. The app is primed.
Every role scored against your compass, with status tracking across your pipeline.
Six dimensions, fit summary, signal words, gap analysis. Calibrate scores you disagree with.
When the posting asks for something your Book doesn't evidence, the app doesn't stop at "not a match." It turns a rejection risk into a weekend plan.
Strong fits return nothing. That silence is the confirmation — you already have the skills this role needs.
Paste a job posting or fetch by URL. Scored in 15-30 seconds.
Claude suggests companies. Each is verified live against its ATS. Track or skip.
Pipeline view per company. Active roles sorted by urgency, watchlist by category.
Every API call, its cost, and what it was for. Export to CSV.
Work in progress. Cost estimates are computed from token counts — they haven't been reconciled against actual Anthropic billing yet. Treat as directional.
The Book — career history, projects, skills, writing rules, sensitive context.
The Compass — scoring rubric. Signal words, red flags, role tiers, compensation floor.
The Playbook — resume rules, cover letter rules, banned words, version strategies.
All three are plain markdown. Edit them in any text editor to customize how Claude scores and writes for you.
Change a rule, and the next score reflects it. No retraining, no waiting.
Use Revise in the sidebar to regenerate from the intake, or edit the files by hand for full control.
Choose your AI model, tune signal words, reload source files, and access Revise.
Generate cover letters and resume summaries, then download as clean PDFs. Arial font, single column, no graphics. Exactly what ATS scanners want.
Available on the materials page after generating content for a role.
Local-first. Your data stays on your machine. Runs on the Anthropic API or Claude Code.