Respec your career for the AI era.
A free tutor that runs inside your coding agent, with its source code on GitHub. It reads your background, finds the AI-era role that fits you in current job-market data, and teaches it hands-on until you can prove it.
Start in two commands
git clone https://github.com/getrespec/getrespeccd getrespec && claudethen type /respec start
A frontend developer's build, before and after a respec
Skill points: 41
- React: 9/10
- TypeScript: 8/10
- Accessibility: 7/10
- Design systems: 7/10
- CSS architecture: 6/10
- Legacy frameworks: 4/10
- Streaming UI: 0/10
- Tool use: 0/10
- Evals: 0/10
- RAG: 0/10
Showing the Senior Frontend Developer build.
How it works
Open the folder with your agent and type one command. Respec takes it from there, one step at a time.
Profile
Respec reads your CV or asks a few questions, then runs a 15-minute placement check so it knows what to skip.
Role discovery
It matches you to two or three AI-era roles using sourced job-market data, and says why each one fits or doesn't.
Curriculum
A personal plan of 12–20 modules for the role you pick, tied to what real job postings ask for.
The loop
Every module is a lesson, a quiz, a review of what you missed, an assignment, and an evaluation.
Dashboard
Your progress, weak spots, and level, from Apprentice to Master, on a page you can open any time.
Learn what to use, where, and when
Knowing what RAG is won't get you hired. Knowing when it beats a longer context window will. Every module teaches decisions, and grades them.
Lesson
Runnable examples, a comparison table of alternatives, and a bridge from what you already know.
Quiz
Scenario questions: in this situation, which would you choose and why?
Review
Missed something? It comes back with a different example, never the same explanation twice.
Assignment
Build the thing. Respec gives hints, not solutions.
Evaluation
A rubric, evidence from your code and your app's real LLM traffic, and two questions about your choices.
Weak spots return in later quizzes until you get them right twice.
Roles, chosen from job-market data
Respec ships with a sourced map of AI-era roles, and your agent can refresh it from the web. Each figure keeps its source and date.
| Role | Good fit if you | Latest signal |
|---|---|---|
| AI-fluent developer | want to keep your title and add AI skills; the shortest path | 75% of US tech postings ask for AI skillsDice, Jul 2026 |
| AI Engineer | like building products on top of LLMs, end to end | Fastest-growing role on LinkedIn's 2026 list, USLinkedIn, Jan 2026 |
| AI Product Engineer | prefer the interface layer: streaming, tool calls, agent UX | No standard title yet; posted as Frontend or Applied AI EngineerGreatFrontEnd, Oct 2026 |
| Forward-Deployed Engineer | enjoy shipping inside a customer's systems and talking to them | Postings up more than 1000% year over year, Jan–Aug 2026Lightcast via Fortune, Sep 2026 |
| AI Consultant | would rather design how a company uses AI than write most of the code | Second fastest-growing role on LinkedIn's 2026 list, USLinkedIn, Jan 2026 |
| ML Engineer | want to train and fine-tune models, and enjoy the math | Median total compensation $282,750, USLevels.fyi, Oct 2026 |
These are different kinds of metrics, so they can't be compared with each other. Respec never mixes them when it recommends a role.
Runs inside your agent
Claude Code, Codex, or Cursor. Respec is a folder of instructions, so there's nothing to install and no new subscription.
Speaks your language
Type /respec start, then write in Turkish, Spanish, or German. Lessons, quizzes, and feedback follow you.
Keeps your data on your machine
Your CV, progress, and assignments stay in a local folder that's never committed. Usage logs record tokens and latency, never prompts or keys.
See where you stand
One command builds a dashboard of your curriculum, scores, topics due for review, and what your assignments cost in tokens. It updates after every step.
/respec dashboard

Pricing
The tutor is free, and its source code is on GitHub. Paid plans add things that need a server.
Tutor
Free
Available now
- Profile, role discovery, and a personal curriculum
- Lessons, quizzes, reviews, assignments, and evaluation
- Local dashboard and LLM usage logs
- Bring your own agent and API key, or use a local model
Pro
Monthly
Coming soon
- A sandbox API for assignments, so you don't need your own API account
- Evaluations that read your app's real LLM traffic
- Progress sync and a web dashboard
- Market data and role packs kept current for you
Certification
One-time
Coming soon
- Capstone evaluated on our servers
- Questions about your design decisions
- A public page employers can use to verify it
Prices will be announced at launch.
Questions
Do I need to pay for anything?
No. You need a coding agent you already use (Claude Code, Codex, or Cursor). Assignments that call an LLM need an API key, a provider's free tier, or a local model through Ollama.
Which agents does it work with?
In Claude Code you use /respec commands. In Codex and Cursor you type respec start, respec next, and so on. Recommended model settings come with the repo.
Is it only for frontend developers?
No. Backend, fullstack, mobile, and data developers all start from the same placement check, and role discovery works from your own background.
How long does a transition take?
It depends on the role and your weekly hours. Sources in the market data put a frontend-to-AI-Engineer move at six to twelve months. Respec estimates yours when it builds your curriculum.
Where does my data go?
Nowhere. Everything Respec writes about you lives in the learner folder on your computer, which git ignores. The usage proxy records model names, token counts, and latency only.
Can I use Respec at work?
Yes. Its license (FSL-1.1) lets you use, change, and share Respec for any purpose, including at work or to train your team. The one thing it doesn't allow is offering Respec, or something built from it, as a competing commercial product. Two years after each release, that version becomes available under Apache 2.0 too.
Is my progress a certificate?
No. Your progress lives in plain files on your computer that you can read and change, so it's your learning record, not proof. Certification is a separate assessment on our servers: your capstone, questions about your own design decisions, and a short task you haven't seen before. Each certificate has a public page employers can use to verify it.