echoby voirlabs

Open source · MIT · MCP

Applications in
your own voice.

Proposals, cover letters and applications written by the AI you already use, grounded in your real work and checked against how you actually write.

Works with Claude, ChatGPT or your own API key.

Your knowledge base✳Claude & ChatGPT over MCP✳Bring your own key✳Style check✳Self-hosted

01 / Features

Keep the AI.
Lose the AI voice.

The model does the writing, on your own plan or key. Echo makes sure it knows your work and sounds like you.

01

One knowledge base about you

Profile, experience, projects and real samples of your writing. The relevant parts get pulled in for each brief.

02

Works inside Claude and ChatGPT

An MCP connector with OAuth. Paste a brief into a chat, and the assistant drafts, checks and saves it to your account.

03

Or bring your own key

The browser extension writes on any page with Anthropic or OpenAI and inserts into the field. Your key stays in your browser.

04

A style check, not a guess

Your style is measured from your samples: openings, length, lists, punctuation, AI tells. Checks are deterministic, and the server never calls a model.

02 / How it works

Set it up once.
Use it everywhere.

Nothing is ever submitted for you. You stay the one who clicks send.

  1. Add yourself

    Your profile and three to five pieces you're proud of.

  2. Connect a writer

    Claude or ChatGPT over MCP, or an API key in the extension.

  3. Paste the brief

    Write, check against your style, save.

  4. Insert it

    Click a field on any page. On Upwork it fills the apply form too.

03 / Self-host

Your writing stays on your server.

A Next.js app, Postgres and a Chrome extension you build for your own URL.

  • OAuth 2.1 with PKCE for MCP clients
  • No model calls on the server, so no API bill
  • The extension reads a page only when you open it
Read the quickstart
git clone https://github.com/voirlabs/echo.git
cd echo && pnpm install
docker compose up -d     # Postgres
cp apps/web/.env.example apps/web/.env
pnpm db:setup
pnpm dev                   # http://localhost:3000