Skip to main content

Quickstart

By the end of this tutorial Matilda will be running on your machine, you will be signed in to the dashboard as an Owner, and you will have chatted with her in the browser.

Prerequisites​

  • Python 3.11 or newer
  • Pango, which the PDF renderer needs
  • A Kimi API key from platform.kimi.ai

Install Pango on macOS with Homebrew:

brew install pango

On Ubuntu:

sudo apt-get install -y libpango-1.0-0 libpangoft2-1.0-0 fonts-dejavu-core

1. Install and configure​

Clone the matilda-webapp repository, create a virtual environment, and install dependencies.

python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env

Open .env and set the four values below. Everything else can stay commented out.

KIMI_API_KEY=your-kimi-key
MATILDA_DASHBOARD_OWNER_EMAILS=you@example.com
MATILDA_DASHBOARD_SEED_PASSWORD=change-me-on-first-login
MATILDA_DASHBOARD_SECRET=a-long-random-string
tip

Matilda boots without any of these. Without a Kimi key she replies with a "getting set up" notice. Without an owner email you will have no way to sign in to the dashboard, so set at least that.

2. Start Matilda​

python run.py

Set MATILDA_RELOAD=1 in the environment if you want the server to restart on code changes. Matilda listens on http://localhost:8000 and writes state to ./data.

3. Sign in to the dashboard​

Open http://localhost:8000/dashboard/login and sign in with your owner email and the seed password.

Because this is a first login, Matilda redirects you to your profile with a banner asking you to set a new password. Choose one and save. You land on the Portfolio Overview.

4. Confirm the model connection​

Open Setup in the sidebar. Find the Kimi node and click Test. Matilda sends a one-word completion to the model and reports success or the error it received.

5. Chat with Matilda​

Open http://localhost:8000/login in a second tab. Enter a phone number and an email address and click Send code.

Without a Resend key configured, the six-digit code is printed to the server log instead of emailed. Copy it from the terminal, paste it in, and click Verify.

You are now in the web chat. Say hello. Matilda introduces herself and asks your name. Answer a few questions and watch the lead appear under Leads in the dashboard with its qualification bar filling in.

Next steps​