Getting started

You'll have your first snapshot in minutes. This guide covers the three things that get you there: create a free account, look around and add your first brand, then upgrade when you're ready to pull live data.

No credit card and no license key to start. Signup is free.

Step 1: Create your free account

Go to app.geotally.ai/signup.

Enter your full name (optional), a work email, and a password (8+ characters). That's it — no license key, no card. The signup screen says it plainly: free to start.

Then check your email. We send a verification link; click it to activate the account. Until you confirm, the account isn't active, so don't skip this step. If the email doesn't land within a few minutes, check spam.

The email + password you set here is your sign-in credential from now on. A session holds for 7 days; after that, sign in again at app.geotally.ai/login.

Already bought through getbdshield.com? You have a license key (BDSH-XXXX-XXXX-XXXX) and you don't enter it at signup anymore. On the login screen, use "Have a license key but no password yet? Sign in with license →", sign in with the key once, then set a password from Account → Security. Your tier comes from the license. See Account management for how license-key billing works.

Step 2: Explore, then create your first brand

You land on the free tier. A read-only demo brand is pre-loaded so you can see what populated data looks like — snapshots, mention rates, share-of-voice, the lot. Click around it.

When you're ready, create your own. The free tier gives you 1 brand with up to 5 prompts. The free tier can't run the engines — that's gated behind a paid tier — so you're setting up now and pulling live data after you upgrade.

The dashboard shows an empty state with a single button: New brand.

A brand has two required fields and one optional one.

Brand name (required)

The exact name you want the system to look for in AI responses. For most companies this is your product name as it appears on your homepage — "Stripe", "Notion", "Linear". Capitalization matters less than spelling; we normalize case.

If your brand name is a common word (e.g., "Apple" is also a fruit), expect some noisy matches — the parser leans on your domain and the surrounding answer text to disambiguate, but it isn't perfect. A more distinctive prompt set helps more than anything: questions where your category is explicit produce cleaner matches.

Domain (required)

Your primary domain — stripe.com, notion.so. We use this to:

Match cited URLs that point to your site (counted as "self-cited")
Disambiguate your brand from name collisions
Build the share-of-voice baseline

Use the apex domain. Don't include www. or https://. If you have multiple domains (acquisitions, regional sites), add the one your customers know best — citation matching runs against this single domain today.

Competitors (optional but recommended)

Add 1-5 named competitors as Name + Domain pairs. Examples for a CRM:

Salesforce, salesforce.com
HubSpot, hubspot.com
Pipedrive, pipedrive.com

Competitors unlock share-of-voice. Without them, the dashboard tells you whether you were mentioned. With them, it tells you whether you were mentioned more or less than the alternatives the engine is also considering. That second number is the one that drives decisions.

You can add or remove competitors later. Changes apply from the next run forward — earlier snapshots keep the competitor list they were captured with, so add your main rivals early if share-of-voice history matters to you.

Write your prompts while you're here. Click Add prompt to write your first one. Read Writing good prompts before you write more than one — a bad prompt set produces noise that's hard to clean up later.

Step 3: Upgrade to run live data

The engines are gated on the free tier. To pull real data, upgrade in-app.

Open the Billing page at /app/billing. You'll see plan cards with Choose Starter / Choose Growth / Choose Agency buttons. Pick the tier that matches your scale:

Starter — 1 brand, 5 prompts, weekly cadence (one run per prompt, spread across the week). Right for solo founders or anyone validating whether this category of tool earns its keep.
Growth — 3 brands, 15 prompts, each prompt runs 3× a week. The default for in-house marketing teams.
Agency — 5 brands, 50 prompts, daily runs, white-label PDFs, priority support. Built for agencies tracking client brands.

The button opens secure card checkout. After you subscribe, your tier unlocks, the engines turn on, and the read-only demo brand is cleared. You manage the subscription from the same Billing page — Manage subscription opens our billing portal to update your card, change plan, cancel, or download invoices.

Billing is monthly. Annual plans aren't available yet — when they land, expect the standard save-two-months shape. You can switch tiers anytime — see Account management for upgrade and downgrade rules.

There's no fixed money-back window. If Geotally isn't earning its keep, email support@geotally.ai — we handle refunds case by case and we're reasonable about it.

What happens when you run a prompt

Once you're on a paid tier with a prompt saved, you have two options: wait for the next scheduled run (06:00 UTC — daily on Agency, 3× a week on Growth, weekly on Starter), or click Run now to trigger an immediate run from your tier's run budget.

A manual run hits every engine the brand tracks in parallel — all 6 by default, or the subset you've enabled in the brand's engine settings. The dashboard shows engine progress live: ChatGPT done, Claude in flight, Gemini done, Perplexity in flight, Copilot done, Google AI Overviews in flight. Total time runs 15-60 seconds depending on engine response time.

When all 6 engines finish, the snapshot finalizes. You'll see:

Mention rate for that prompt
Average rank when mentioned
Top cited URLs
Competitor share-of-voice

If an engine fails (rate limit, timeout, content-policy refusal), the snapshot still finalizes with the engines that succeeded. We mark the failed engine on the snapshot so you know the sample is partial.

Two things to do in the first week

Tune your prompts

Your first prompts will be wrong. That's normal. After 3-5 days of data:

Drop prompts that no engine answers (too vague, or too niche)
Drop prompts that always mention you (you don't need to track them — they're solved)
Drop prompts that never mention you and never mention a competitor either (the question doesn't exist in the engines' worldview)
Keep prompts where rank or share-of-voice is moving

This filtering matters because your prompt slots are finite. On Starter you have 5 slots. Spending one on a question no engine answers is a waste.

Turn on the weekly digest

In Account → Notifications, the weekly digest is on by default. Leave it on for the first month. It'll land Mondays UTC with week-over-week deltas and the biggest movers. If you find yourself ignoring it, turn it off — but most users say it's the moment they actually look at the data.