> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getvector.me/llms.txt
> Use this file to discover all available pages before exploring further.

# Start a project from a template

> Create a financial model from a built-in template, pick the starting numbers, and keep it on the latest version of its template.

<Callout icon="wrench" color="#F59E0B">
  **This area is still in development.** The product marks Modeling as **WIP**: screens, labels and behavior described here can change, and some steps may not work yet.
</Callout>

Building a financial model from a blank page is slow, and it's easy to miss a cost line or a pricing mechanic that changes the whole picture. In Vector you start from a template instead: a working model with realistic assumptions already in place, which you adjust to your own product.

Each model is a project — a separate experiment with its own numbers, [actuals](/modeling/actuals-and-forecast) and [snapshots](/modeling/snapshots-and-accuracy).

## Choose a template

Open **Templates** to see what's available. Every [workspace](/account-and-team/workspaces) gets the same built-in templates, including ones added after the workspace was created.

* **Notetaker** — a per-seat SaaS model: acquisition funnel, customer base, team, flat subscription pricing and standard SaaS costs.
* **AI SaaS — Token Economics** — the same funnel, customer base, team and cash mechanics, built around an AI product. It replaces flat per-seat pricing with a usage-quota tier, and counts the AI infrastructure around your model calls, not only the inference itself, in the cost of serving customers.

Templates you've made yourself are marked **custom**, next to the **built-in** ones. You can delete a custom template; built-in ones can't be deleted.

## Create a project

<Steps>
  <Step title="Start a new project">
    Open **Modeling** and click **+ New project**.
  </Step>

  <Step title="Name it">
    Give it a **Name**, and under **Description** say what you're testing — for example, a new pricing idea.
  </Step>

  <Step title="Pick the template and the starting numbers">
    Choose the **Template (model)**. If it offers more than one **Starting preset** — a named set of starting numbers — pick one. The template decides how the model is built; the preset decides which numbers it opens with. The built-in templates have a single preset, so the choice doesn't appear for them.
  </Step>

  <Step title="Create">
    Click **Create**. The project opens straight away, fully calculated from the template's numbers.
  </Step>
</Steps>

Nothing is left blank: every assumption, chart and KPI shows a value from the start, and you change any of them afterwards. See [Assumptions and dashboard](/modeling/assumptions-and-dashboard).

## Keep a project on the latest template version

Templates are versioned, and a project stays on the version it was created from — editing a template never changes existing projects. The project shows which version it's on under its name.

When a newer version is out, a banner at the top of the project says so. **Upgrade to v…** moves the project to it and carries over your values and actuals, with nothing lost.

## Save your model as a template

Once you've shaped a project into something you'd reuse, click **Save as template** at the top of the project, and give the template a name and description. It joins the list, ready for the next project — with your project's numbers as a starting preset.

## Compare two approaches

Both built-in templates share the same funnel and team structure. Create one project from each on comparable numbers to see, side by side, how AI-specific costs and pricing change the economics compared with a plain per-seat SaaS model. See [Scenarios](/modeling/scenarios).

## Where to go next

<Columns cols={2}>
  <Card title="Set up your workspace" icon="arrow-left" href="/getting-started/set-up-your-workspace">
    Where modeling sits in the setup checklist.
  </Card>

  <Card title="Record actuals and forecast from them" icon="arrow-right" href="/modeling/actuals-and-forecast">
    Bringing real numbers into the model.
  </Card>
</Columns>