Pricing and tokens
Pricing and tokens
What a test, a research question and the other paid tools cost, what the plans include, and what costs nothing.
How paying works
A token is the unit everything here is paid in. A test, a research question and the other paid tools each take a set number of tokens from your balance. Every test costs 300 tokens, whatever its length and however many shares it covers. The Strategy Agent and the Research Agent show the cost before anything runs.
Verify your email to get 3 free tests (1,000 tokens). The 100 tokens left after the 3 free tests pay for 10 look-ups, or 2 written answers.
The price list
| What you run | Tokens |
|---|---|
| A research question answered by looking figures up (a price, a company figure, a dividend, an interest rate) | 10 |
| A research question an AI writes up from those figures, or from a web search with its sources listed | 40 |
| Deep research: several look-ups, written up with sources | 120 |
| A research playbook: several deep-research sections | 250 |
| One test, whatever its length and however many shares it covers | 300 |
| A goal search of 24 candidates, its usual size | 2,100 |
A question the Research Agent cannot answer costs nothing. A goal search costs more than the free 1,000 tokens, so it needs a plan. A few other tools have their own prices, shown before they run: a company Deep Dive, the Prove it check, and refreshing a saved strategy.
Plans
Plans are sized in tests. These are read from the plan table each time this page is opened:
| Plan | Price | What it pays for |
|---|---|---|
| Starter | £15 a month | 25 tests a month. 7,500 tokens a month, at 300 tokens a test. |
| Pro | £40 a month | 50 tests a month. 15,000 tokens a month, at 300 tokens a test. |
| Enterprise | £85 a month | 150 tests a month. 45,000 tokens a month, at 300 tokens a test. |
Pricing has the same plans with their buttons.
What costs nothing
- Questions the Strategy Agent asks you before a test.
- A test that is refused, or whose rule never buys anything: it is refunded automatically. So is a run that fails.
- Tracking a rule with pretend money, and opening a sealed test.
- Cost estimates in your AI app.
- The stock screener, paper trading with pretend money, your saved strategies and their results, Market Pulse, the evidence explorer and the Education Centre.
When your tokens run out
- Nothing runs that your balance cannot pay for. The run is refused first, and nothing is charged.
- On a plan, the next monthly payment adds that plan's tokens again. Tokens you have not used stay in your account.
- Changing to another plan is done by email for now. Billing tells you where to write.
Related
- PricingPricingThe plans, with their buttons.
- Feature guidesThe Strategy AgentThe page where you type an idea and get a test: what it reads, what it asks, what it charges, and what you can do with a result.
- Feature guidesThe Research AgentAsk about a company, a fund, a market or the economy. Each answer says where its figures came from, and when.
- FAQ and troubleshootingFAQ and troubleshootingAnswers to the questions people ask most: why the agent asked a question, why a test was refused, why a result changed, and more.
These pages describe how Quantis Trade works and what its tests show about the past. They are not forecasts and not advice.