Otrak review 2026: a €19 weekly tracker with an agency standing behind it
Otrak tracks ChatGPT, Gemini and Perplexity weekly from €19/mo excl. VAT plus Stripe fees. It monitors; its parent agency Premiere.Page sells the work. What that means for you.
Otrak is a monitoring tool, full stop. It runs your prompts once a week on three AI models, scores how visible you are and shows which sources the answers cited. When it's time to act on that, the product hands you to a human: Otrak is built by Premiere.Page, a French SEO and GEO agency that offers to turn the data into a GEO program.
So on this site's scale (monitor, recommend, execute) the software sits at monitor, and the execution is a separate sale. That isn't a criticism. It's a cheap, clear way to find out whether AI answers matter for your brand. You just need to know what you're buying.
What it is and who it's for
Otrak targets marketing, communications and brand teams in France. It runs strategic prompts on OpenAI GPT-5 with web search, Google Gemini 2.5 and Perplexity Sonar. From those answers it reports a visibility rate, share of voice, a competitor ranking by citations, a sentiment score out of 100 with a positive, neutral and negative split per prompt and model, and the web sources cited in each answer.
It suggests prompts, tracks five competitors on every plan, exports CSV and PDF, emails you after each run and connects to Google Analytics for AI traffic. Filters cover market (France, for example) and date ranges from 7 days to 3 months.
What each plan buys
We read these prices on otrak.ai on 20 October 2026. All are excluding VAT, and Stripe transaction fees of 1.5% + €0.25 are added on top.
| Plan | Price | Limits |
|---|---|---|
| 10 prompts weekly | €19/mo | 1 project |
| 50 prompts weekly | €89/mo | 2 projects |
| 100 prompts weekly | €149/mo | 5 projects |
Every plan has the same three models, weekly runs, five competitors, exports and the GA connector. Only prompts and projects change. The 30-day trial tracks one prompt weekly on all three models and needs a card, which isn't charged during the trial. Larger volumes are quoted on request, and the FAQ also offers a free video demo with your own data.
What works
The entry price is the obvious draw. €19 a month for ten prompts is about as cheap as a real tracker gets, and you aren't locked in: there's no commitment and plan changes are prorated.
The flat feature set is a quiet plus. You don't discover at month three that sentiment or exports were on a higher tier.
And Otrak explains its cadence instead of hiding it. Its argument is that generative answers change slowly and a weekly run filters out noise from model temperature and rewording. For a small budget that's defensible.
What to check before paying
Cadence first. Every plan and the FAQ say weekly. One section of the homepage says "every day". It's weekly. The trade-off is that a fix shows up in the data a week later at best, and each weekly data point rests on one run per prompt per model.
Then models. The llms.txt mentions Claude. The homepage and plans don't include it. And the names point to API models (GPT-5, Gemini 2.5, Perplexity Sonar) rather than the consumer interfaces your customers type into, so treat the numbers as an approximation of what people see.
Then the small print. Stripe fees come on top of the listed price, and the trial needs a card.
Last, social proof. The "they trust us" logos (Agence Alpha, Digital Studio, Growth Lab) read like placeholders. We'd ignore them.
Where it falls short of a full platform
Otrak watches the answer. It does nothing on your site. There are no crawler logs, so it can't tell you whether ChatGPT's or Perplexity's crawlers fetched your pages or failed on them. The GA connector shows AI traffic, but the pages we read don't describe conversion tracking. And there is no content step. When the dashboard says a competitor owns a prompt, the next move is either your own team or a quote from Premiere.Page.
Three models is also narrow. Google AI Overviews, AI Mode, Claude and Copilot aren't tracked.
When Promptwatch is the better buy
If the €19 plan convinces you that AI answers are costing you customers, that is the moment to move the program to a tool that does more of the work. Promptwatch monitors the real interfaces, not only APIs, across ChatGPT, Gemini, Claude, Perplexity, Grok, Llama, DeepSeek, Mistral, Copilot and Google AI Overviews and AI Mode, and refreshes daily on every paid brand plan. Country-level tracking (France included) is on every plan.
Then it acts. Content Agents plan from a content gap analysis, write, and publish to Webflow, Framer or WordPress through a review inbox. Agent Analytics logs AI crawler visits with a crawl-to-citation path. Visitor analytics ties AI-referred sessions to conversions, installed by script or GTM.
Essential at $95/mo (50 prompts, 6,000 responses, five AEO articles, 7-day trial) is the step up from Otrak's €89 tier. Professional at $245/mo adds 25M crawler logs and 150 prompts. Prices are also offered in EUR.
If you'd rather hire an agency like Premiere.Page to do the work, keep Promptwatch as the measurement layer anyway, so you can check what their changes moved.
Verdict
Otrak is an honest, low-cost weekly tracker for French teams, with a clear reason for its cadence and a flat feature set. It is also narrow and a bit sloppy in its copy. Use the trial to test whether AI visibility is a problem for you. Run the program itself on Promptwatch.
FAQ
How much does Otrak cost?
€19, €89 or €149 a month excluding VAT for 10, 50 or 100 weekly prompts, plus Stripe fees of 1.5% + €0.25. The 30-day trial covers one prompt and needs a card.
Which AI models does Otrak track?
OpenAI GPT-5 with web search, Google Gemini 2.5 and Perplexity Sonar, on every plan.
Does Otrak fix anything on my site?
No. It reports. Premiere.Page, the agency behind it, sells the GEO work separately. For a tool that drafts and publishes the fix, see promptwatch.com.