Notra: AI visibility tracking for your brand

What is Notra used for?

Notra is an AI visibility tracking and GEO platform for teams that want to know whether AI assistants recommend their brand. Give it the questions a potential customer would ask, and it checks selected engines, saves their answers and shows who appears instead when your business is missing.

For example, a software company can monitor a question such as “Which tools help a small team manage product updates?” It can inspect the named brands and cited pages, then decide what information its own site needs. This makes Notra useful for founders, marketers, SEO teams and agencies with capacity to act on the findings. The service is a web application operated by Notra, Inc.

Run a first scan that you can learn from

1. Create a workspace and enter your public website. Review the brand name, aliases and suggested questions rather than accepting an inaccurate description.
2. Add relevant competitors. Under Prompts, add questions manually, import a CSV or use suggestions from a connected Google Search Console account.
3. Choose a plan that includes AI answers, then set your engines, languages and scan schedule. A project can track up to four languages; its prompt language occupies one slot.
4. Open Overview and select Run Scan. Inspect the actual answers and sources for important questions, alongside mention rate, position, sentiment and share of voice.

Keep the question set and settings consistent when comparing runs. If a known brand mention is missed, check its aliases first. The official quickstart explains the setup and troubleshooting.

Separate AI recommendations from crawler visits

The visibility reports show how often the selected engines name your brand and how it compares with tracked competitors. Full answers and source pages help explain a result; a position inside an AI answer is a different measurement from a conventional search ranking.

AI Traffic adds another view. The server-side @usenotra/geo SDK can run as middleware or a proxy on supported frameworks including Next.js, Nuxt, TanStack Start, Astro and SvelteKit. It separates training crawls, search-index crawls, fetches made while answering and referral clicks.

A crawler request is not proof that the final answer cited you. Read the saved answer and citations before claiming a recommendation. Homepage dashboard numbers are illustrative sample data, not measured results for your business.

Turn missing mentions into an editorial task

Content Gaps identifies questions where competitors appear and your brand is absent, then ranks opportunities. Choose a gap and a guide, listicle or comparison format. Notra uses your brand identity, sitemap and competitors to plan a brief; approve the brief before the draft opens in Content.

Studio also turns authorized GitHub, Linear or Slack activity into material such as changelogs, blog posts and social updates. Review the selected source activity and brand voice, then check the draft's claims before publication. API, CLI and MCP access can connect these tasks to a developer or agent workflow, subject to authentication and scopes.

Understand the free plan and answer allowance

Checked on October 29, 2026: the free plan provides AI traffic analytics and agent feedback for up to 1 million events per month. Ongoing prompt tracking and the writer are paid features.

The public monthly USD plans list Starter at $100 with 2,000 AI answers, 5 million traffic events, 10 long-form posts and one project; Growth at $250 with 6,000 answers, 15 million events, 25 posts and three projects; and Scale at $550 with 12,000 answers, 30 million events, 50 posts and ten projects. Enterprise is quoted separately. Annual billing is also offered.

One answer means one prompt × one engine × one language × one scan. Ten prompts across three engines and two languages use 60 answers each scan. Confirm current allowances, add-ons, taxes and billing terms on the official pricing page.

Check data handling before connecting sources

The privacy policy says AI traffic logs can include paths, user agents, referrers, approximate country and language. IP addresses are used transiently and are not stored with those logs. This does not mean the whole service stores no data.

Zero data retention for model-provider scans is an optional 20% surcharge on Starter, Growth and Scale, and is included on Enterprise. It restricts the eligible model set unless you approve exceptions. The terms say Notra does not train AI models on connected integration data and that users retain their content rights; generated content still needs human review. Read the privacy policy and terms before connecting sensitive business material.

Frequently asked questions

Can I track AI answers for free?

The advertised free plan covers traffic analytics and agent feedback. Do not assume it includes continuous prompt monitoring or paid writing allowances.

Which engines and languages can I check?

Core web-grounded scans include ChatGPT, Claude, Gemini and Perplexity. Other engines may have no web search. The live catalog controls availability, and each project can select up to four supported scan languages.

Does Notra keep my prompts out of model-provider storage?

That depends on the ZDR setting and eligible providers. ZDR is a plan option, not a blanket claim that Notra itself retains no account, content or traffic data.

Official resources and community

• Documentation: setup, metrics and developer access.
• GitHub organization: official developer projects.
• Discord community: community discussions.
• Notra on X: product announcements.
• Contact Notra: support and sales inquiries.