For traders who want fewer mistakes, not more tools

Less data. Better decisions.

Behavioural finance and the information you actually need.

Right information. Right mindset. That's what makes an honest trader.

No card. Free forever.

app.samosfi.com / terminal
pulse
At a glance
Wed, 28 May · 21:04
Today's read
Your top bias: FOMO / herd 72 / 100
Symbol24h
BTC67,234.12+1.24%
ETH3,421.08−0.41%
SOL156.42+2.11%
what makes us different

Four lines we won't cross.

01

AI questions. Never instructs.

Other crypto tools sell signals — "buy this, sell that". We never tell you what to do. Our AI is built structurally to ask, mirror, and remind. This is both philosophical and legal: we sit on the right side of the investment-advice line, and you stay in the driver's seat.

02

Methodology open. No black box.

Every signal, every behavioural score, every guardrail pattern is visible in our code and documentation. Three filter layers protect every AI output. No "trust us, we're smart". The exact logic that decides what reaches you is published.

03

Behavioural finance integrated.

Most retail losses come from behaviour, not from missing information. We mirror your traps — loss aversion, FOMO, disposition effect — and put them in front of you when you're about to trade. Academic foundation: Kahneman 1979, Shefrin-Statman 1985, Barber-Odean 2001.

04

Free for the test + the watchlist.

The bias test, the bias library, and a 5-coin watchlist are free forever. Premium adds the coach and behavioural tracking. No paywalls on the parts that teach. No ads anywhere.

what's inside

Three layers. No clutter.

The information that matters and the mindset that survives drawdowns.

layer 1 · live

Know your traps

12 questions. Plain English. Your top three biases, named.

FOMO
Disposition

Take the test

?
layer 2

Pause before you act

A 3-beat check-in: thesis, context, exit plan. Mindset, not advice.

Is this a thesis, or a reaction?···
layer 3

See yourself, in numbers

Your trades stop being claims. They become evidence.

Win hold
2.1d
Loss hold
9.4d
how will you use this

Three paths in.

New to crypto trading

You're curious. You've made a few trades or none. You don't want to learn by losing your savings.

  1. Bias library — read the 8 traps before they happen
  2. Take the bias test — 3 minutes, know your tendencies
  3. Add 2-3 coins to watchlist — observe without acting
  4. When you do trade — run the check-in. Discipline first.

Active trader, want discipline

You've been trading for a while. You know your wins and losses. You want a structure that resists emotional decisions.

  1. Run the bias test — confirm what you suspect about yourself
  2. Set up watchlist + signals — eliminate the rest of the noise
  3. Pre-trade check-in on every position — the friction that compounds
  4. Reflect monthly — winners-vs-losers held days, consistency rate

Burnt by FOMO, want to slow down

You chased something. You held a bag too long. You're not sure why you keep doing it. You're ready to look at the pattern.

  1. Bias library — the chapter on FOMO / herding is where to start
  2. The bias test will probably confirm it — that's OK, naming helps
  3. Coach pre-trade: forces the pause when the move feels urgent
  4. Reflect view shows the held-time gap — proof, not lecture
how it works

Semi-automated. Human-reviewed.

Let's be honest about what runs on AI and what runs through human eyes.

automated

Data + signals + coach

  • · Price data polled every 60 seconds from public exchange feeds
  • · Behavioural scoring runs deterministically from your bias test answers
  • · Coach responses come from a free-tier language model behind three guardrail layers (input filter, system prompt, output check)
  • · Critical-signals feed is filtered, de-duplicated, and ticker-tagged programmatically
human-reviewed

Editorial + safety

  • · Editor pieces are written and signed; no AI ghostwriting
  • · Coach guardrails are pull-request reviewed; pattern changes are compliance events
  • · Red-team test suite runs in CI; adversarial inputs are checked before deploys
  • · Honest blind spot: the AI can still be wrong. If you see a slip, the feedback button is one click away

We don't pretend to be fully autonomous. AI handles scale, humans handle judgement.

pricing

Free where it teaches. Paid where it tracks.

Free
€0
Forever
  • Bias test + 8-bias library
  • Watchlist up to 5 coins
  • Pulse view + signals feed
Premiumrecommended
€9.99/mo
Cancel any time · MVP test mode
  • 3-beat check-in coach
  • Live prices + behavioural tracking
  • Reflect: your trade evidence

Premium is in test mode during MVP. No real payments are taken yet — we're finding the first 50 EU traders who care about behaviour before flipping the switch.

honest questions

FAQ.

Are you a financial advisor or investment service?

No. We don't and won't recommend buy / sell / hold actions, price targets, or directional predictions. The AI is built structurally to refuse those. This is both a product choice and a regulatory position.

How many users do you have?

Pre-revenue. We're looking for the first 50 EU crypto traders who genuinely care about behaviour, not just signals. We'd rather earn 50 users who use it weekly than 50,000 who download and forget.

Where does my data go?

Bias test results stay in your browser unless you sign up. Watchlist, check-ins, and your behavioural profile (post sign-up) live in encrypted EU storage. We never sell data and we don't run ads.

Why is the coach free if AI calls cost money?

For MVP we use free-tier models (Groq Llama 3.3 70B + Llama 3.1 8B audit). Per-user cost is negligible. When traffic grows, Premium subscribers will subsidise heavier model use; Free will keep the bias test, library, and watchlist.

Is this for the US?

Marketing focus is EU — we're based in Dublin. The product is in English and accessible globally; we don't actively solicit US users.

Why behavioural finance instead of signals or backtests?

Because most retail losses don't come from missing information. They come from acting against the plan, holding losers, chasing winners, FOMO. Backtests don't fix any of that. Naming the trap before you fall into it does.