How can we help?

Setting up your first brand takes about 3 minutes. Open Workspace → General → Add brand and you'll be guided through 5 steps.

1. Brand basics

Add your brand name exactly as customers write it. The domain (e.g. loopstack.io) is used for linked-mention detection. Aliases matter most for brands with informal spellings or handles — add as many variants as you can think of.

2. Description

Write 2-3 sentences about what you do. This is used to power AI phrase suggestions in step 3. The more specific, the better the suggestions.

3. Competitors

Add up to 6 competitors manually, or let AI suggest based on your description. You can adjust the list anytime from Competitors in the sidebar.

Good phrases match how real buyers ask LLMs questions. We classify them into 5 types:

• Branded — "is Loopstack any good?" — high intent, easy wins
• Comparison — "Linear vs Loopstack" — most valuable for competitive positioning
• Category — "best PM tool for design teams" — hardest to win, biggest reach
• Feature — "PM with Figma integration" — narrow but high-converting
• Use case — "how to manage cross-team sprints" — top of funnel

Aim for a mix. Most teams start with 25 phrases and grow to 100-200 within 3 months.

The visibility score is a single 0-100 number summarizing how AI assistants talk about your brand.

What goes in

• Presence (40%) — did your brand appear in the answer?
• Rank (25%) — what position were you mentioned at?
• Sentiment (20%) — positive, neutral, or negative tone?
• Share of voice (15%) — how much of the answer was about you vs. competitors?

We show 7-day rolling averages by default because LLM responses drift even at temperature=0. Raw daily scores are available in the trends view.

Go to Workspace → Notifications → Channels → Slack and click "Connect Slack workspace". You'll be redirected to Slack to grant chat:write permission.

After connecting, pick a channel and configure which events fire there from the notification matrix. We recommend starting with the "Just the important stuff" preset.

Every webhook is signed with X-Glow-Signature: t=<timestamp>,v1=<hex-hmac>. Verify by computing HMAC-SHA256 over <timestamp>.<raw-body> with your signing secret.

Node.js example

const crypto = require('crypto');
const sig = req.headers['x-glow-signature'];
const [t, v1] = sig.split(',').map(s => s.split('=')[1]);
const expected = crypto.createHmac('sha256', SECRET)
  .update(t + '.' + req.rawBody).digest('hex');
if (expected !== v1) return res.status(401).send();

Reject signatures older than 5 minutes to prevent replay attacks.

Notifications go to people. Pretty-formatted alerts in Slack, email, or the in-app bell. Configured by marketers.

Webhooks go to systems. Signed JSON POSTs to your backend, Zapier, or n8n. Configured by developers.

Both can fire on the same events — the choice is who you want to reach.

If you're not satisfied within 30 days of your first paid subscription, email billing@llmglow.com and we'll refund you in full.

The guarantee applies to first-time subscriptions only, once per customer. Refunds are processed within 5 business days to your original payment method.

No. Your prompts and brand data are never used to train, fine-tune, or improve any AI model — ours or a third party's.

Technically: we use LLM provider endpoints contractually committed to no-training (OpenAI Enterprise zero-data-retention, Anthropic API, Google Vertex with data-residency). Where supported, we pass X-No-Training hints in headers.

This is also in our Privacy Policy section 4.