Two ways to run it. One question decides which.
Do you want to bring your own model or agent, or would you rather not think about models at all? That single choice sets what you pay, who you pay it to, and how many analyses your plan includes. This page lays both out with the numbers.
Reading time: 6 min · Last reviewed: 2026-09
Available now in the browser, and in the extension this autumn.
The AI analysis, the three webQsee AI strengths, Deep Inspection and the new plans are live today in webQsee web at view.webqsee.com: open a recording there, analyse it, and buy a plan or a booster if you want one. The Chrome and Edge extensions are in review and follow this autumn. Your recordings, your gallery and your account are the same in both, so nothing you do now has to be done again.
The short answer
There is no per-analysis fee to webQsee in either case. What your plan sets is how many analyses you run a month, because the work around the model is ours: turning a recording into something a model can read, scoring the moments, doing the coverage arithmetic, wiring every finding back to the analyzer row it rests on, and keeping the answer with the recording. With BYO AI you pay whoever runs it. With webQsee AI the model is on us, which is why that allowance is the smaller of the two.
Who pays for the model.
It is the only real difference between the two, and everything else follows from it.
BYO AI
Point the quick analysis at any OpenAI-compatible endpoint, or let Deep Inspection start the coding agent you already have. The excerpt goes straight from your browser to the address you configured. Point it at a model on your own machine and nothing leaves the device at all.
- Your API key never leaves the extension's own storage
- A local model means the recording never leaves your machine
- The only variant that can run Deep Inspection
- Follow-up questions are unlimited and never counted
Analyses included
Pro is €59.90 and Studio €360 a year, billed annually. VAT included.
One pot for both variants: the quick analysis and Deep Inspection draw from the same number, so sixty means sixty analyses and not sixty of each. The free plan's ten are given once rather than every month; Studio's are shared across the whole team. Deep Inspection needs Pro, or any BYO AI booster.
webQsee AI
No key, no endpoint, no model to pick. Press Analyze and read the answer. The excerpt goes to webQsee and is answered by a model we pay for, so the analyses come out of your plan rather than out of an account you have to open somewhere else. There are three strengths of it, and you pick one per analysis: Standard, which your plan includes every month, and Smart and Genius, which are bought as boosters.
- Nothing to set up, and nothing to pay a third party
- Masking is on by default, because the destination is not your machine
- We keep counters, an analysis id and the model's name, never the question or the answer
- Quick analysis only, because Deep Inspection runs on your machine by definition
Analyses included
Pro is €59.90 and Studio €360 a year, billed annually. VAT included.
These are the Standard numbers, the strength your plan includes. Follow-up questions inside one analysis are capped here, because they cost us tokens too: 3 on Free, 8 on Pro, 10 on the team plans. A follow-up replays the whole conversation, which is why the cap on the two stronger strengths is tighter still, 6 and 3. The free plan gets 5 Standard analyses on the house - once per account rather than every month, so that trying the hosted analysis costs nothing and needs no key - and no monthly allowance after those. It also starts with 10 BYO AI analyses, once, and after those a BYO AI booster is the cheapest way on: we buy no model tokens there, and you buy none either if you run a model on your own machine. The strength boosters are on sale to it like to anyone else.
The three strengths
| Strength | When to pick it | Booster, incl. VAT |
|---|---|---|
| Standard | Included in every paid plan, and the quickest of the three. Handles most recordings on its own. | 50 for €11.90 |
| Smart | One step up, for when a Standard answer did not settle the question. You can escalate an analysis you already started. | 10 for €15.90 |
| Genius | The largest model we host, for when you want the most we can put behind a question. | 8 for €18.90 |
How long they take. Standard answers in a few seconds. Smart and Genius take longer, usually well under two minutes. We quote those two together on purpose: we measured them over the same recordings and they overlap almost entirely, and on two of five recordings Genius was faster than Smart or level with it. Any recording can be quicker or slower than this, so the window counts the seconds as they pass rather than predicting them. Follow-up questions inside an analysis are usually quicker than the first answer, on every strength, though not on every single run.
How much stronger, honestly. We ranked all three blind against each other on five recordings, with a judge from none of their families. The order holds on average, and it is not a cliff: Standard placed first on one of the five, ahead of both paid strengths, and Smart and Genius are close enough over five recordings that we do not claim one beats the other. None of the three invented a finding. So the honest way to use them is to start on Standard and escalate the same question if the answer does not settle it, rather than to guess in advance which recording needs which strength.
Which model sits behind each strength is our decision and it changes as better ones appear, which is exactly why we sell a strength rather than a model name. Every one of them runs inside the EU on a route that keeps nothing: no retention, and never trained on. The analysis window tells you afterwards which model answered, and a saved analysis keeps that on record.
Both sets of bars are drawn on the same scale, 0 to 360, so the two lanes can be compared directly. Free's two numbers are one-time starter allowances rather than monthly ones, and Studio's are shared by the whole team rather than given to each member; the bar shows the number, the label beside it says which kind it is. The webQsee AI bars are the Standard strength; the other two are bought rather than included, so they have no monthly number to draw.
Most of what you do never counts.
An analysis is counted when a question is answered, not when a button is pressed. The bottom half of this table matters more than the top half.
| What you do | Counts |
|---|---|
| Asking a new question about a recording | one |
| Comparing two recordings with each other | one, not two |
| Starting Deep Inspection from the extension | one |
| Follow-up questions inside the same analysis | — never* |
| Carrying on in the terminal after Deep Inspection started | — never |
| Opening a saved analysis again, days later | — never |
| An attempt that failed: network error, wrong key, your provider rate-limiting you | — never |
| Retrying the same question after a timeout | — never |
| Pressing Analyze and then closing the dialog | — never |
| Your own agent reaching recordings through the helper's MCP server | — never |
Running wqs-helper analyze on the command line | — never |
* Unlimited with BYO AI. With webQsee AI a single analysis allows 3 to 10 follow-ups on Standard depending on your plan, and 6 or 3 on Smart and Genius, because we are paying for those answers and a follow-up replays the whole conversation.
Offline, it still runs
If our backend cannot be reached, the analysis goes ahead anyway. Up to three are remembered and counted the next time you are online. Working on a train is a delay, not a discount.
A paid allowance resets. A free one does not
A paid plan's allowance is per calendar month, UTC. The free plan's ten BYO AI and five webQsee AI analyses are the exception: they are given once, for the life of the account, and do not come back on the first. On a Studio the monthly allowance belongs to the team rather than to one person, so all five members draw from the same three hundred and sixty and any one of them could spend them all, exactly as the cloud-item cap already works. A booster bought for that team goes into the same pot, because the two halves of one purse have to live in the same place.
Five boosters: one for yours, three for ours, and the pack.
Running out is an offer rather than an error: the panel tells you where you stand and what tops it up. Boosters are one-time purchases that stay valid for a year, and every price here includes VAT. One bought for yourself is yours; one bought for a team is the team's, and the surface you buy from says which it is before you pay.
75 analyses for €11.90
The cheapest thing we sell, at about 16 cents an analysis, and it opens Deep Inspection for as long as it lasts. Valid for a year. The €5.90 Booster Pack also carries 25 of these, alongside the longer videos, the extra cloud items and the 24-hour share links.
It costs something even though you bring the model, because the model is not what you are paying for: we turn the recording into something a model can read, run the MCP round trip and build the report, and that work is the same size whoever pays for the tokens. If you need analyses every month, a Pro subscription at €59.90 a year is the cheaper answer, and we would rather say so than sell you packs.
See what else is in the packOne booster per strength
Each strength is its own product, because a booster has to be countable in the unit it was sold in. The €5.90 pack carries a handful of each as a taste; these are for when you know which strength you want. All three stay valid for a year.
- Standard: 50 analyses for €11.90
- Smart: 10 analyses for €15.90
- Genius: 8 analyses for €18.90
Each strength has its own balance and they never lend to each other, which is the point: "four Genius analyses left" is something you can act on. Your plan's monthly Standard allowance is always spent first, then Standard boosters, and among boosters the one expiring soonest. You never lose an allowance by holding a booster.
Go to pricingThe AI numbers, per plan.
The AI rows in this table are live in the browser at view.webqsee.com today.Extension: autumn
| AI analysis | Free | Pro | Studio | Enterprise |
|---|---|---|---|---|
| Analyses, BYO AI (your own model or agent; per month on a paid plan) | 10 once, not monthly | 60 | 360 (shared across the whole team) | unlimited (each member) |
| Analyses, webQsee AI Standard (nothing to configure; per month on a paid plan) | 5 once, to claim | 20 | 120 (shared across the whole team) | 250 (each member) |
| webQsee AI Smart and Genius | Bought as boosters, on any plan including Free. No plan includes them. | |||
| Follow-up questions, BYO AI | unlimited | unlimited | unlimited | unlimited |
| Follow-up questions per webQsee AI Standard analysis | 3 | 8 | 10 | 10 |
| Follow-up questions, Smart / Genius | 6 and 3, whatever the plan, because there a follow-up costs us about as much again as the analysis | |||
| Deep Inspection (agent on your own machine, counts against BYO AI) | with a BYO AI booster | |||
| Comparing two recordings | ||||
| Your own agent over MCP (never counted) | ||||
| Account required | Yes, for every counted analysis: the allowance is kept against it | |||
| Switching strength inside one analysis | Counts. Each strength has its own balance, so the first question you ask on a new one starts a new analysis there. Coming back to a strength that already answered is free again. The analysis window says so before you send. | |||
| Allowance resets | First of the month, UTC – except the free plan's starter allowance, which is given once and does not come back. On Studio the allowance belongs to the team and is shared; on every other plan it is counted against the team you created, so each member has their own | |||
Anonymous capture stays free and needs no account. It is the analysis that needs one, the same way the error-detection rules already do.
Which one should you use?
Just starting
Use webQsee AI Standard if you are on a paid plan: there is nothing to configure and you can judge the feature on its answers instead of on a setup form. On the free plan, point the quick analysis at a model on your own machine and it costs nobody anything. Reach for Smart or Genius when Standard has already answered and the answer did not settle it, the panel offers to re-ask the same question one strength up. That is a new analysis in the stronger strength's balance, and the window says so before it sends.
Recordings you cannot send anywhere
Use BYO AI with a model running locally. The excerpt never leaves the machine, and masking is yours to switch off because there is nothing to protect it from. This is the only configuration where no data travels at all.
You already have a coding agent
Use Deep Inspection, and map your project so the answer can reach the source behind the behaviour. Whatever you already pay your agent covers the model; your plan only sets how many inspections you start.
Where each one sends your recording
The destination is written above the question box every single time you ask, because "runs on my machine" and "is posted to a provider" are two clicks apart and look identical otherwise. The privacy page lists what leaves your device for every variant, including what masking does and deliberately does not hide. Read what it sends, and where
Stop guessing. Start webQseeing.
Add webQsee to Chrome or Edge in one click. Recording and replay are free, no signup needed. Upgrade when you need AI analyses, cloud sharing, S3 storage and Pro-grade tooling.
Works in Chrome 103+, Edge 103+ and most Chromium-based browsers. Install instructions.