Home

Where you stand and what to do next, built from everything you have fed in.

Today

Loading...

Inputs checklist

Reads the record, the market, your training and the pipeline, and gives three things to do this week. Runs on its own every seven days.

Skills record

Every skill has a level and a trust grade, backed by evidence. The CV is one export of this.

Saving your CV, role notes or chat notes, or syncing GitHub, drafts evidence and skills here on its own. Build record re-runs that by hand. Nothing lands until you accept it, because an accepted item becomes a claim with a trust grade. Public exports hide private items and show client work as outcome only.

Skills

Nothing yet.

Trust: 1 self declared (no evidence) · 2 evidence linked (a role, project, talk, publication, repo or course) · 3 assessed (question bank, or a passed cert without a link) · 4 verified (credential link or endorsement). Skills not evidenced for over a year are marked stale.

Target roles

For each target role on your profile: what the saved postings for it ask for, how much your record covers at level 2 or more, what is missing, and a ready date from your tracker target dates and pack forecasts. No model call.

Visibility

Talks, publications and repos by year, from your evidence. The plan suggests twelve pieces for the next quarter from your own work.

Evidence

Add evidence by hand

Nothing yet.

CV

CV: one export of the record

Written from the skills record and your material, fact checked against both. Keep it current for recruiters who still expect one.

Tailor to a saved job

Add postings on the Jobs and market page first.

Learn

Practice packs with original questions, the certs worth doing, your training tracker, and where AI fits your field.

Practice

Building uses web search to read the vendor's skills outline. Check the objectives in Creator studio before you rely on them.

No packs yet.

Certs and training

Build a learning path

Certification or exam

Target date

Hours/week

Training tracker

Cert, course or project

Type

Target date

Hours/week

AI in my field

AI in my field

Where AI applies to your field and your role, decided from your own material: what to learn, what to build, where it fits the work you do now, and the guardrails. Uses web search.

Extra focus optional, e.g. "workshop delivery" or "partner reporting"

Creator studio

Pick a pack, or build one on the Practice page.

Market

What the market asks for, from live samples and the postings you save, matched against your record.

Market demand

Role families one per line, defaults to your target roles

Region

Trends

No samples yet.

Samples

None yet.

Saved postings

Add a posting

URL optional: works on pages that expose the text, otherwise paste it Posting text

PipelineSaved postings (0)

Each posting is broken into skills, tools and certs and matched to your record. Switch on job search mode on the Profile page to track stages, follow ups and replies.

None yet.

What the market asks for

Amber bars are items that do not appear anywhere in your CV, certs, role description or public repos. It is a text match, so treat it as a prompt to check.

Build

Proof on GitHub: what your repos already show, and what to build next for your gaps and the market.

GitHub

GitHub username

Private repo detail sent to Claude

Repos

Not synced yet.

What to build

Chat

A coach that learns how your role actually works. Save what it learned and every analysis reads it.

What Career Compass knows from chat

Notes only record what you said. Edit freely: this is exactly what gets sent with your other analyses.

Profile

Your inputs, the modes, and the settings.

Current CV

CV text paste or upload, you can edit it here

What you do and what you want

What we do in my current job team, clients, services, scope, how work gets done, who you report to Roles and directions I am interested in titles, sectors, size of company, remote or hybrid, anything you would rule out Interests topics you like working on, what you would learn for fun

Certifications, training and time

Certifications held name, year, expiry if any Training in progress or planned Constraints location, travel, notice period, salary floor, anything the plans must respect Hours per week I can spend learning

Settings

Your keys

This copy of Career Compass runs entirely in your browser. Your key and your data are stored here and go nowhere else; the only calls that leave this page go from your browser to Anthropic and, if you add a token, to GitHub. Use an Anthropic key with a spend limit, because anything saved in a browser can be read by the extensions you have installed.

Anthropic API key GitHub token optional: a fine grained token with read only Contents and Metadata for private repos

Model Thinking

Modes

Both off by default. Each one adds pages and controls.

Job search mode pipeline stages and follow ups on saved postings, search agents, interview prep, offers and contacts, CV tailoring Creator mode the creator studio: author packs, generate and verify questions, learner reports, pack export

Writing style rules

Applied to any text drafted for you: CV lines, pitches, messages.

Reviewer pass on CV drafts a second model call audits grounding and tightens weak lines before you see the draft. Doubles the cost of those tasks. Appearance stored in this browser only Monorepos to look inside comma separated repo names. Each top level folder is treated as a project.

Your data

Everything lives in the data folder next to the server. Nothing is stored anywhere else.

Search agents

Agents that find postings for you, scored against your record, within the run cap.

Weekly review on the Brief page, every 7 days while the server is up

Monthly run cap

This month

What I am looking for

Titles

Seniority

Locations

Remote rules

Salary floor

Sectors

Deal breakers Titles to exclude from board scans comma separated

Searches and scoring read what you kept and dismissed before (with your dismiss reasons), so each run leans toward what you keep.

Company boards

Reads the public job APIs of companies you care about (Greenhouse, Lever, Ashby, Workable). No AI cost to scan. Matching titles land in Candidates unscored, then you score them with AI when you want to.

Candidates

None yet.

Notifications

None.

Run log

None.

Interview prep

Stories from your evidence and a mock interview for any saved posting.

Pick a saved posting to aim at, or leave it on your target roles. Save postings on the Jobs pipeline page.

Debrief a real interview

Right after the interview, type what they asked and how it went. You get the competencies tested, gaps exposed, stories to sharpen, and a follow up note.

Mock interview

Pick a posting and press Start. Answer out loud or type, as you would in the room.

Offers and contacts

Offers compared on total value and what you care about; the people in your search.

Offers

Add an offer

Company

Title

Base per year

Bonus amount, or percent if ticked

%

Pension employer contribution, amount or percent

%

Equity annualised value as you see it

Sign on

Other cash car, allowances

Holiday days

Location and mode

Growth 1 to 5

Role fit 1 to 5

Flexibility 1 to 5

Stability 1 to 5

Deadline

Currency

Notes

No offers yet.

Contacts

No contacts yet.