Free Poker Solver Options That Are Actually Free

Free solver options fall into three groups: open-source CFR engines you run yourself, the free tiers of subscription training sites, and free calculators that answer most solver questions without any solve at all. What none of them hands over is a large presolved library with a trainer attached, which is the thing most players are actually shopping for. Here is what each free option does, exactly where it stops, and how to build a study routine that never crosses the paywall.

The three things people call a free solver

Open-source CFR engines. Free to download, free to run, free forever. You supply the ranges, the bet tree, the board and the electricity. Nothing is precomputed, so every answer costs you wall-clock time on your own machine.

Free tiers of paid products. A subscription site gives you a slice — a few spots per day, flop nodes only, one game format — and keeps the library behind the paywall. Genuinely useful, deliberately narrow.

Free tools that are not solvers. Equity calculators, pot-odds calculators, ICM calculators, preflop charts. None of them runs counterfactual regret minimisation, and for most of the questions you would put to a solver they answer faster and just as correctly.

The third group is the one players underrate, and it is where the free lane is widest.

The options side by side

Option	What it costs	What you get	Where it stops
Open-source CFR engine	Nothing, plus your hours	Any spot you can define yourself	No trainer, no ranges, no grading
Free tier of a trainer	Nothing, until the daily cap	Real solver output, packaged	Spot limits, shallow trees, one format
Equity calculator	Nothing	Hand vs hand, hand vs range	No betting, no strategy
Pot-odds and ICM calculators	Nothing	Break-even prices, bubble pressure	One number at a time
Preflop charts	Nothing	Opening and defending ranges	Fixed stack depth and format
Desktop solver	One-off licence	Unlimited custom trees	Steep learning curve, your hardware
Subscription library	Monthly	Huge presolved library, trainer	Recurring cost, fixed trees

Read down the last column rather than the first. The cost differences are obvious; the ceilings are what decide whether free is enough for you this year.

Running an open-source engine yourself

This is the purest free option, and the least popular for good reason. You get a CFR engine and nothing else: no ranges, no board library, no interface worth the name, no feedback on whether your input made sense.

Three costs are hidden in that word "free".

You build the tree. Every bet size at every node is a choice you make. Give the engine three sizings on each street and it finishes quickly. Give it seven and the tree explodes, memory fills, and a single flop can run for hours.

You supply the ranges. The solver answers the question you asked. Feed it a button opening range that is far too tight and it will hand you a perfectly optimal strategy against a range nobody has. Garbage ranges in, confident nonsense out.

You interpret the output. Raw strategy grids do not tell you what to fix. Reading them is a separate skill from playing, and it is the reason so many downloaded solvers get opened twice.

What free tiers of paid products actually give up

Every commercial trainer with a free tier draws the line in roughly the same three places.

Volume. A handful of hands or spots per day. Enough to see the interface work, not enough to drill a leak until it sticks.

Depth. Flop strategy is cheap to serve and turn and river subgames are not, so free access usually thins out fast after the flop. That matters, because the flop is the street where most players are already competent.

Custom work. Entering your own ranges, node locking an opponent's tendency, or solving a stack depth the library does not carry is the feature that costs the most to run, so it is nearly always the first thing gated. On Sharkling the free tier covers solver-backed drills and calculators. Premium includes the covered cash library and covered-hand review; Premium+ adds a monthly Custom Solve quota and node-locking for supported spots.

None of that is a trick. Presolved libraries cost money to compute and to store, and any product giving one away without limit would not exist for long.

The free tools that replace most solver queries

Before you fight an engine, notice what you were going to ask it. Most solver questions at low and mid stakes reduce to one of three numbers, and all three are free.

How often am I ahead? That is a free equity calculator, not a solve. A♠K♠ against QQ is roughly 43/57 — a slight dog, not the coin flip folklore claims. AKs against 22 is close to 50/50. A flush draw with two cards to come gets there about 35% of the time.

What price am I being offered? A pot odds calculator turns any bet into a break-even percentage instantly. Facing $50 into $100 you risk $50 to win $150, so you need 25%. Compare that against the equity number and the decision is made.

What should I be opening? Work from free preflop charts by position, which settle the highest-frequency decision in the game. UTG opens roughly 15-17% of hands in 6-max and the button roughly 40-45%, and playing those two seats identically costs more per month than any subscription.

That covers preflop ranges, calling prices and hand-versus-range equity — the categories where low-stakes money actually leaks. A solver is for what remains after those are clean.

A study routine that stays free

1. Fix preflop first. Learn opening ranges for all six seats, then defending ranges from the blinds. This is free, it is finite, and it is worth more bb/100 than any postflop work you could do instead.
2. Guess, then check. Mark two or three hands per session, write down the price and your equity estimate from memory, then verify with a calculator afterwards. The gap between guess and answer is your actual leak list.
3. Drill inside the free cap. Whatever daily allowance a trainer gives you, spend it all on one spot type until the pattern is automatic, rather than sampling random spots for variety.
4. Only then run a solve. When you can already name both ranges in a spot, a solver tells you something you could not have worked out. Before that, it is expensive decoration.

Followed properly, that routine takes months to exhaust. Players who go straight to a downloaded engine usually skip step one entirely, which is why solvers built for beginners put charts and graded drills in front of raw strategy grids.

When paying starts to make sense

Run the arithmetic instead of the vibes. If you play 20,000 hands a month at $0.25/$0.50, a single bb/100 of improvement is worth 200 big blinds, or $100. Any tool that reliably buys you that much clears a subscription several times over, and any tool that does not is a hobby purchase.

The honest triggers for paying are specific:

• Your preflop game is solid and your losses are concentrated on turns and rivers.
• You need a spot the free library does not carry — an unusual stack depth, a multiway pot, a specific ICM situation.
• You want to node lock a population tendency and see how to exploit it, which no free tier offers.
• Your time is worth more than the licence. Waiting three hours for a flop to solve is only free if your hours are.

If none of those apply yet, the paid tier will not fix anything the free lane has not already fixed. When they do apply, the full solver comparison covers how the desktop and subscription options differ in tree control, learning curve and presolved coverage.

The short version: free is enough for longer than most players believe, and the moment it stops being enough is a moment you can name.