Naruvi is an open-source, self-hosted career workspace. It keeps your job postings, fit checks, CVs, cover letters, emails, interviews and offers in one place, with an AI companion (an otter called Naruvi) that helps you take the next step.

Yes. Naruvi is licensed under Apache 2.0. You run it on your own computer or server. If you use a hosted model provider, you pay that provider directly; with a local model like Ollama there’s no model bill at all.

Which AI models does it work with?

Anthropic natively, any OpenAI-compatible API, and local models through Ollama. You bring your own key, and you can switch at any time in Settings.

Will it invent experience I don’t have?

No. Drafts are built only from facts you’ve verified in your profile, and document checks flag claims you can’t back up. Evidence before eloquence.

Where does my data live?

In your own PostgreSQL database and asset storage. Saved model credentials are encrypted at rest with AES-256-GCM. Your model provider only receives the context a workflow needs.

You need Docker with Compose v2, Git and OpenSSL. A guided installer sets Naruvi up locally, on your own domain with automatic HTTPS, or behind Coolify, Traefik or nginx.

Does it support internships and apprenticeships?

Yes. Naruvi handles all kinds of roles, including full-time, part-time, internships and alternance contracts.

Open source · Self-hosted · AI-native

A calmer way to find your next role.

Naruvi saves the roles, checks your fit, drafts your CV and letters from facts you’ve verified, and keeps every reply and interview in one workspace you run yourself.

Install Naruvi free See how it works

Never invents experience
Every draft starts from facts you’ve verified.
Runs on your machine
Your CVs, mail and contacts stay on your server.
Free and open source
Apache 2.0. Bring your own model, or run one locally.

The Naruvi dashboard: an offer to review, applications that need attention, and upcoming next steps

Naruvi1 offer on the table

That’s real progress. When you’re ready, I can help you compare it.

One application, from saved to sent.

Naruvi does the reading, matching and drafting. You make every decision.

Naruvi

Saved. I pulled out 9 requirements. Check your fit whenever you’re ready.

Less admin. More applications you believe in.

A job search is mostly bookkeeping. Naruvi takes the bookkeeping, so your evenings go to the roles worth your time.

The usual way: Postings scattered across tabs, bookmarks and alert emails
With Naruvi: Every role on one board, requirements already pulled out
The usual way: Guessing whether you’re a fit, then writing anyway
With Naruvi: A fit check against your verified facts, gaps named first
The usual way: The twelfth cover letter, from a blank page
With Naruvi: A CV and letter drafted for this job from your own facts
The usual way: Replies buried in your inbox, follow-ups forgotten
With Naruvi: Emails matched to applications, next steps on your dashboard
The usual way: Interviews tracked in your head
With Naruvi: Interviews and holds in your calendar through a private feed

Honest about you. Private by default.

Every claim about you is backed by a fact you verified. Everything Naruvi knows stays on a server you run.

Fit check · Senior Frontend Engineer

Strong match

5+ years building React applications
Copperline, Tidewater Studio · 6 years
Proven
TypeScript in production
Verified in 4 roles
Proven
Design systems at scale
Led a component library; scale not stated
Partly
Mentoring junior engineers
Mentored 3 engineers, 2024
Proven
Build features end to end with design
What you’d do in the role
Your call

Drafts only use what’s proven

Perks and instructions are kept apart from real requirements, and document checks flag any claim you can’t back up.

Your server

Stays with you

Your workspace
Profile, verified facts and every application, in PostgreSQL
Your documents
CVs and letters in local asset storage
Naruvi
The app and its background worker

Only the context a task needs, only to what you connect

Your model
Any provider, or local with Ollama
Your pick
Your mailbox
To match replies to applications
Optional
Notifications
Reminders through ntfy
Optional

Encrypted at rest

Saved credentials use AES-256-GCM. Run a local model and nothing reaches a model provider.

Ask Naruvi

Ctrl J

What needs me this week?

Three things, most urgent first:

Follow up with Brindle Learning
1 day overdue
Overdue
Technical interview at Northwind Labs
Wednesday, 10:00
Tidewater Studio documents are ready
Send or set a date

Nothing changes until you confirm.

Lost track? Just ask.

Ask about your search in plain words. Naruvi answers from your own workspace and proposes the next action. Nothing changes until you say so.

Try a question:

Running in a few minutes.

You need Docker with Compose v2, Git and OpenSSL. Read the script, run it, and finish setup from the private link it prints.

Terminal

$ curl -fsSLo install-naruvi.sh https://raw.githubusercontent.com/Eljakani/naruvi/main/install.sh
$ less install-naruvi.sh   # read it first
$ bash install-naruvi.sh

Or let your agent install it

Naruvi is AI-native. Paste this into Claude Code, Codex or Cursor and it handles the rest.

The installer asks where Naruvi runs:

On this computer
The quickest start. Ports bind to localhost, so only your machine can reach it.
On your domain
Point a domain at your server and the installer sets up automatic HTTPS.
Behind your proxy
Already run Coolify, Traefik or nginx? Naruvi sits behind it.

Works withAnthropicOpenAI-compatibleOllamaIMAP / SMTPGmail OAuthMicrosoft OAuthiCalendar feedntfyWhisperREST APIMCP serverBrowser extensionPostgreSQLpgvectorDocker

Ask your AI about Naruvi.

Naruvi is AI-native, so the quickest guide is the assistant you already use. Pick a question and open it with the prompt already written.

I'm looking at Naruvi (https://github.com/Eljakani/naruvi), a self-hosted, open-source job search workspace. Read the README and docs, then walk me through using it day to day: saving roles, checking my fit, drafting a CV and cover letter, and tracking replies and interviews. Keep it practical.

Ask Claude Ask ChatGPT Ask Perplexity

Questions, answered plainly.

Something missing? Open an issue.