Minecraft Librarian Trade Calculator

Work out how many lectern resets a librarian takes to offer the enchanted book you want, what it costs in emeralds before and after a zombie cure, and whether an enchanting table beats the grind. Java and Bedrock, from the game's own 26.2 data files.

Minecraft 26.2 (the 1.21 line, renamed) · data checked September 4, 2026

What book do you want

Edition Java Bedrock

Book you're hunting

All 40 enchantments a librarian can sell at 26.2. Each one is exactly 1 in 40: a trade draws with equal chances, so Silk Touch is no rarer than Efficiency here — unlike at an enchanting table.

Level

Max level only, any level, or one exact numeral. A numeral the enchantment does not have reads “Not possible” rather than quietly rounding down.

Max emeralds

The most you would hand over for the book, after the cure discount. 64 accepts anything — a librarian never asks for more.

Your setup

I'll cure this villager from a zombie

A cure knocks 20 emeralds off every book permanently, which widens the prices worth accepting. One cure is the maximum — a second adds nothing.

Seconds per lectern reset

Break the lectern, place it again, read the trade. 8 seconds is a steady pace; nothing publishes a measured figure, so this one is your assumption to set.

What the lectern will cost you

Lectern resets, on average

1.67% chance on each fresh librarian

9 times in 10, by

138 resets

One hunt in ten runs longer than this.

Chance per reset

1.67%

1 in 60

Time at the lectern

8 min

60 resets at 8 s each

What you'll pay

24 emeralds

10–38 emeralds, plus 1 book

At an enchanting table

Never

No table can roll it.

Table or librarian?

The librarian is the only way. Mending is a treasure enchantment: no enchanting table can ever roll it, at any bookshelf count. 60 resets on average.

How many resets before you have it

Chance you have it by then	Lectern resets	Time
Half the time 50%	42	6 min
3 in 4 75%	83	11 min
9 in 10 90%	138	18 min
99 in 100 99%	275	37 min

The average is 60 resets, but half of all hunts finish inside 42 and one in ten drags past 138. That gap is the difference between “about 60” and “I'll be done in 60”.

Chance of a book trade at each villager level

Villager level	Chance of a book trade
Novice	67% 0.67 book trades rolled by here
Apprentice	67% 1.33 book trades rolled by here
Journeyman	67% 2 book trades rolled by here
Expert	50% 2.5 book trades rolled by here
Master	0% 2.5 book trades rolled by here

A Java librarian taken all the way to Master has had about 2.5 book trades rolled — but only the first one can be rerolled. The moment you trade once the villager locks and the lectern trick stops working. Levelling adds new chances at a book; it never improves a book you already have.

What the book costs, and what a cure takes off

The villager	Emeralds off	You pay
Never traded with reputation 0	none	10–38
Traded with a lot reputation 25	−5	5–33
Cured, later the same day reputation 100	−20	1–18
Cured just now reputation 125	−25	1–13
Cured, then traded to the cap reputation 150	−30	1–8

Prices for Mending . The listed price is capped at 64 emeralds before any discount, which is why 64 turns up far more often than any other figure on a level V book. Hero of the Village is the one discount missing from this table: it comes off at the counter, changes no reroll odds and wears off.

Sizing a trading hall

Every book, once

257 librarians

Every book at max level

933 librarians

Every book at max level for 1 emerald

8,695 librarians

Coupon-collector figures for Java: every librarian is one draw at all 40 books at once, so the tail is long — the last handful of enchantments take longer to turn up than the first thirty put together.

What this calculator does not model

The Villager Trade Rebalance experiment. It ships as a separate bundled datapack that replaces the librarian's trade tables, so it is not the default game. Under it each biome gets a three-enchantment pool and a guaranteed Master book — a swamp librarian is guaranteed Mending — and every number on this page would be wrong.
Your world's actual sequence. Since Java 26.1, rerolls are generated from the world seed, so a seed-aware predictor can tell you what the next reset gives. This page computes the odds across worlds, which is a different and complementary answer.
Hero of the Village. It is a percentage off at the counter, applied outside the reputation term, and it changes no reroll odds — so it moves the price you pay and nothing else on this page.
Combining books on an anvil, and the prior-work penalty that makes the fifth combine cost 39 levels.

Trade pools, the price formula and the reputation values are read from the vanilla datapack at 26.2 and the Minecraft Wiki. Minecraft © Mojang Studios.

Updated: September 5, 2026

Minecraft librarian trade calculator. Lectern resets, odds and emerald price for any enchanted book.

A Minecraft librarian trade calculator works out how many lectern resets it takes before a fresh librarian offers the enchanted book you are hunting. It reads the trade tables from Minecraft 26.2 (the 1.21 line, renamed), so it also gives the emerald price, how many resets nine hunts in ten finish inside, and whether an enchanting table would get there sooner.

What is a Minecraft librarian trade calculator?

A Minecraft librarian trade calculator turns the book you are hunting into three numbers: how many lectern resets it takes on average, the chance on any single reset, and what the villager will charge. For Mending on Java with no price limit, those numbers are 60 resets, 1.67% per reset, and a listed price between 10 and 38 emeralds.

Two draws stack to produce that 1.67%. A fresh Java librarian unlocks two novice trades drawn from a pool of three, so 2 in 3 fresh librarians carry an enchanted book at all. The book itself is then picked from the 40 enchantments a librarian is allowed to sell, each with an equal chance. Mending has a single level, so nothing further divides it. Efficiency has five levels and a price window 55 values wide, which is how the same arithmetic reaches 1 in 16,500 for Efficiency V at its cheapest listing.

The trade pools, the price formula and the reputation values behind every figure here are read from Minecraft 26.2 — the 1.21 line under the game's year-based numbering. The pool size is what makes the difference between them matter: 26.x ships 40 tradeable enchantments and 1.21 shipped 39. The familiar "Mending on 1 librarian in 60" is therefore a 26.x figure; against a pool of 39 it would have been 1 in 58.5.

Bedrock fills each level's slots one trade at a time, and the book shares its slot with one other trade, so 1 in 2 fresh Bedrock librarians carry a book instead of 2 in 3. Every reset count on Bedrock is a third longer as a result: Mending averages 80 resets, against 60 on Java. Switching the Edition control moves every card on the page.

An average hides the shape of a hunt, so the spread sits beside it. Half of all Mending hunts finish inside 42 resets, three in four inside 83, nine in ten inside 138, and one in a hundred is still going at 275. That tail is the part worth absorbing before the first reset: 138 resets is an ordinary bad session at a lectern, and reaching it means the odds behaved.

Since Java 26.1 the reroll sequence follows the world seed, so a seed-aware predictor can name the trade your next reset will produce. This page answers the other question — the odds across every world, for the player who does not have their seed to hand — plus the two things a static odds table cannot express: a cap on what you are willing to pay, and a straight comparison against an enchanting table.

How a librarian's book trade gets picked

An offer is the end of a five-step chain, and each step multiplies into the odds on the cards.

1. The level draw. In Java, every villager level unlocks two trades chosen at random from that level's pool. A librarian's pools hold 3, 3, 3, 4 and 2 trades for Novice through Master, and the enchanted book sits in the first four of them. A novice draws 2 of 3, so 2 in 3 fresh librarians offer a book. At Expert the draw is 2 of 4, or 1 in 2. Master has no book at all — its two slots are candles. Bedrock fills slots one at a time from smaller pools, which gives 1 in 2 at the first three levels and 1 in 3 at Expert.

2. The enchantment draw. Once the trade exists, the game picks one enchantment from the tradeable pool with equal chances. At 26.2 that pool holds 40 entries, so every one of them is exactly 1 in 40. Silk Touch is no rarer than Efficiency at a lectern, which inverts the rarity weights that govern an enchanting table. Multiplying the two draws gives the number everyone quotes:

\frac{2}{3} \times \frac{1}{40} = \frac{1}{60} \approx 1.67\%

3. The level roll. A level is then rolled for the enchantment, uniformly across the levels it has. Mending and Silk Touch have one level each and skip this step. Unbreaking has three, so Unbreaking III is a third of Unbreaking's share: 1 in 180. Efficiency has five, putting Efficiency V at 1 in 300.

4. The price roll. The listed price for a level-L book is an integer drawn uniformly from 2 + 3 × L to 6 + 13 × L, which is 5 + 10 × L possible values. Treasure books cost double, and the 64-emerald ceiling is applied after that doubling. Level V spans 17 to 71 before the ceiling, so the seven rolls above 64 are clipped down to it and eight of the 55 end there — which is why 64 turns up eight times more often than any other price on a level V book. Filtering on price is what produces the longest odds on the page: Efficiency V at exactly 17 emeralds needs all four draws to land together.

\frac{2}{3} \times \frac{1}{40} \times \frac{1}{5} \times \frac{1}{55} = \frac{1}{16{,}500}

5. The discounts. Reputation moves the price by 0.2 emeralds per point, and curing the villager from a zombie is worth 100 points, so a cure takes 20 emeralds off permanently. One cure is the maximum — a second adds nothing. Prices never fall below a single emerald.

One rule sits outside the chain and outranks all of it: trading once locks the villager. A completed trade — including the 24-paper trade at the novice tier — freezes the whole list, and breaking the lectern afterwards only removes the workstation. Leveling is the other half of that rule. A Java librarian taken to Master has had about 2.5 book trades rolled across its levels, but only the first can be rerolled, and a new level never upgrades a book already on the list.

How to use this librarian calculator

Six controls sit in two blocks: "What book do you want" holds the edition, the book, the level and the emerald cap, and "Your setup" holds the cure toggle and the seconds per reset.

1. Pick the Edition. Java draws two trades per level from that level's pool; Bedrock fills each slot one trade at a time. Every reset count on the page moves with this control.

2. Choose the book under "Book you're hunting". All 40 enchantments a librarian can sell at 26.2 are there, each equally likely. "Any enchanted book" answers the separate question of how long until a librarian offers a book of any kind, and it hides the level and price controls, because neither applies.

3. Set the Level: max level only, any level, or one exact numeral. A numeral the enchantment does not have reads "Not possible" instead of rounding down without telling you.

4. Set Max emeralds — the most you would hand over, counted after the cure discount. Leaving it at 64 accepts everything, since a librarian never asks for more. Lowering it is the fastest way to turn a short hunt into a long one.

5. Tick the cure toggle if you intend to cure the villager from a zombie. It takes 20 emeralds off every price permanently and rewrites the price cards, while changing no odds at all.

6. Set Seconds per lectern reset. Break the lectern, place it again, read the trade: 8 seconds is a steady pace. No published measurement exists for this, so it stays your assumption to set.

Six cards answer the question. "Lectern resets, on average" is the headline, "9 times in 10, by" is the figure worth planning against, "Chance per reset" restates it as a percentage and a 1-in-N, "Time at the lectern" multiplies the resets by your seconds, "What you'll pay" gives the average and the range, and "At an enchanting table" runs the same target through a table at 15 bookshelves and turns amber when no table can roll the book.

Below the cards, a "Table or librarian?" line states the verdict in a sentence, and three tables cover the spread, the chance of a book trade at each villager level, and the price ladder from an untraded villager to one cured and traded to the cap. A trading-hall block sizes the whole collection, and a closed panel lists what the page does not model.

Java lectern resets, odds and listed price, book by book

Book	Chance per reset	Resets on average	9 hunts in 10 by	Listed price
Any enchanted book	66.7%	1.5	3	5–64
Mending	1.67%	60	138	10–38
Silk Touch	1.67%	60	138	5–19
Fortune III	0.56%	180	414	11–45
Unbreaking III	0.56%	180	414	11–45
Looting III	0.56%	180	414	11–45
Protection IV	0.42%	240	552	14–58
Efficiency V	0.33%	300	690	17–64
Efficiency V at 17 emeralds	0.0061%	16,500	37,992	17

Worked hunts with real numbers

Mending on Java, paying whatever it asks

Inputs: Java, Mending, no cure, Max emeralds at 64. The chance on any fresh librarian is 1 in 60, or 1.67%, and the average hunt is 60 lectern resets for a book listed at 10 to 38 emeralds, averaging 24. The average is the least useful of those figures on its own, because the spread around it is wide:

Chance you have it by then	Lectern resets	Time at 8 s each
Half the time	42	6 min
3 in 4	83	11 min
9 in 10	138	18 min
99 in 100	275	37 min

Half of all hunts finishing by 42 explains why fifty resets feels like the right ballpark. The number to plan against is 138, where nine hunts in ten are done.

The same hunt, with a zombie cure

Curing the villager changes nothing about the odds — it is still 1 in 60 and still 60 resets on average. What it changes is the bill. A cure is worth 100 reputation at 0.2 emeralds a point, so 20 emeralds come off permanently, and the whole price ladder shifts:

The villager	Emeralds off	You pay
Never traded with	none	10–38
Traded with a lot	−5	5–33
Cured, an hour later	−20	1–18
Cured just now	−25	1–13
Cured, then traded to the cap	−30	1–8

A cured Mending book averages 6.4 emeralds. The common claim that a cure makes Mending cost one emerald is true 40% of the time and no more often than that: six of the fifteen prices a Mending book can list are 20 emeralds or less, and only those drop to the one-emerald floor. The other nine land between 2 and 18.

Refusing to pay more than 20 emeralds

Set Max emeralds to 20 with no cure and the hunt goes from 60 resets to 150. The odds per reset fall from 1.67% to 0.67%, nine hunts in ten now need 345 resets instead of 138, and the accepted price averages 15 emeralds across a 10-to-20 range. The cause is arithmetic: a Mending book has fifteen possible listings and only six of them are 20 emeralds or less, so a cap of 20 discards 60% of every successful roll. Cure the villager first and the same cap becomes easy to hit, which is the argument for curing before the hunt starts.

Fortune III — the case where the table wins

Fortune III is the counterintuitive one. A librarian averages 180 lectern resets for it, at 11 to 45 emeralds; an enchanting table at 15 bookshelves, best slot, averages about 5 tries, at 3 experience levels and 3 lapis each. The table wins by a factor of 36, and the same pattern holds across everything a table can reach:

Book	Lectern resets	Tries at a table
Unbreaking III	180	2
Protection IV	240	4
Fortune III	180	5
Silk Touch	60	6
Looting III	180	10
Efficiency V	300	out of reach on diamond, about 12 on gold
Mending	60	never

The librarian earns its keep on the four treasure enchantments it can sell — Mending, Frost Walker, Curse of Binding and Curse of Vanishing — and on level windows a table cannot reach. Every other book in that table is faster at the table, and by a margin wide enough that the emerald price never enters the decision. For the exact per-slot odds on your item, material and bookshelf count, use the Minecraft Enchanting Probability Calculator; for what those levels cost to farm back, the Minecraft XP Calculator.

Efficiency V, and the price of being picky

Efficiency V at any price averages 300 resets — 0.33% per reset, 40 minutes of lectern work at 8 seconds each, and 43.5 emeralds on average across a 17-to-64 range. Hold out for the cheapest listing the game can produce, 17 emeralds, and the average becomes 16,500 resets. That is 0.0061% per reset and around 37 hours of breaking a lectern, and nine hunts in ten would still be going at 37,992 resets. The gap between 300 and 16,500 is the 1-in-55 price roll, and it is the query no static odds table can express, because a table of enchantment odds has nowhere to put your budget. Efficiency V is also one of the few common enchantments where the librarian wins on a diamond pickaxe: enchantability 10 caps the modified level below what the level V window needs, so a diamond pickaxe never rolls it, while a gold one gets there in about 12 tries.

Sizing a trading hall

Every fresh librarian is one draw against all 40 enchantments at once, which makes hall sizing a coupon-collector problem. On Java it takes 257 librarians on average before all 40 enchantments have appeared at least once at any level, 933 before every one has appeared at its highest level, and 8,695 before each of those has also been cheap enough that a cured villager sells it for a single emerald. On Bedrock the three figures are 342, 1,244 and 11,594. The tail dominates all three: the last handful of enchantments take longer to turn up than the first thirty put together, which is why most halls target the six or seven books a player uses and leave the full set alone.

Bedrock: the same hunt, a third longer

Switch the Edition control to Bedrock and Mending goes from 1 in 60 to 1 in 80, or 1.25% per reset. The average hunt becomes 80 resets, half finish inside 56, nine in ten inside 184, and the 99-in-100 ceiling moves to 367. At 8 seconds a reset that is 11 minutes on average and 25 for the nine-in-ten case. Prices do not change between editions — a Mending book still lists at 10 to 38 emeralds — so the whole Bedrock penalty is in the slot mechanics: each level fills its slots one trade at a time and the book shares its slot with one other trade, giving 1 in 2 where Java gives 2 in 3.

Reset strategy, and what to do with the number

Do not trade until you see the book. One completed trade locks a villager's offers forever, lectern or no lectern, and selling it 24 paper at the novice tier counts. Breaking and replacing the lectern costs nothing but the seconds you set in the calculator, so there is no reason to buy anything before the book you want is on the list.
Cure first if you intend to cure at all. Twenty emeralds come off every price permanently, which widens the set of prices worth accepting and is what makes a low emerald cap survivable. One cure is the ceiling; a second adds nothing.
The price cap is the expensive setting. Hunting Mending at any price averages 60 resets. Refusing to pay more than 20 emeralds turns the same hunt into 150, because only 6 of the 15 prices a Mending book can list fall at or under 20. Raise that cap before you conclude the game hates you.
Level the villager after you lock it, never before. A Java librarian taken to Master has about 2.5 book trades rolled across its levels, but only the first can be rerolled with the lectern. Leveling adds fresh chances at a book; it never improves a book already on the list.
Check the table verdict before the first reset. Unbreaking III is 2 tries at a table against 180 lectern resets, Fortune III is 5 against 180. For anything an enchanting table can reach, resetting a lectern is the slower route by an order of magnitude.
Java beats Bedrock by a third. A Java novice draws 2 of 3 trades while a Bedrock one fills its slot from a pool of 2, so every reset count on Bedrock is 4/3 of the Java figure — 80 resets for Mending instead of 60. Set the Edition control before you read anything else.
Set the seconds per reset honestly. Break, place, read is about 8 seconds at a steady pace, which puts the average Mending hunt at 8 minutes and the nine-in-ten case at 18. Walking between two lecterns doubles it, and the time card is the number that decides whether a hunt is worth starting.
Read the nine-in-ten figure, not the average. Half of Mending hunts finish by 42 resets, so reset 100 feels like a curse. One in ten runs past 138 and one in a hundred past 275, which is ordinary. Pick your walk-away point from the 99-in-100 row before you start, and stop arguing with it at reset 90.

What this calculator does not model

Four things move a librarian's trades that this page leaves alone, and each omission is deliberate.

The Villager Trade Rebalance. It ships as a separate bundled data pack that replaces the librarian's trade tables, so it is not the default game — you switch it on under Experiments when creating the world. Under it each biome gets a three-enchantment pool and a guaranteed Master-level book, and a swamp librarian is guaranteed Mending. In such a world every number here would be wrong.

Your world's actual sequence. Java 26.1 moved reroll generation onto the world seed, so a seed-aware predictor can tell you what the next reset will give you. This page computes the odds across worlds, which is a different and complementary answer to the same problem.

Hero of the Village. The effect is a percentage off at the counter, applied outside the reputation term. It changes no reroll odds, so it moves the price you pay and nothing else on this page.

Anvil combining. Two Efficiency IV books make an Efficiency V one, and the prior-work penalty is what makes the fifth combine cost 39 levels. The page prices the book a librarian sells, not the book you assemble.

One further limit is worth naming. Lunge, the spear enchantment added in 1.21.9, is in the tradeable pool and the librarian side handles it at 180 resets on average, but the enchanting-table engine behind the comparison is pinned to Java 1.21 and predates the spear. The card reads "No model yet" instead of carrying an invented number.

Minecraft librarian trades — frequently asked questions

How many times do I have to reroll a librarian for Mending?

60 lectern resets on average in Java, 80 in Bedrock. Half of all hunts finish inside 42 resets and nine in ten inside 138, so 138 is the number worth planning against. At a steady 8 seconds per reset, the average hunt is 8 minutes and the nine-in-ten case is 18.

What are the odds a librarian sells Mending?

1 in 60 on a fresh Java librarian, or 1.67% per lectern reset. On Bedrock it is 1 in 80. The figure is 2 in 3 that the librarian offers any enchanted book, times 1 in 40 that the book is Mending.

Can I get Mending from an enchanting table?

No. Mending is a treasure enchantment, and no enchanting table can roll one at any bookshelf count. Villager trades, fishing and chest loot are the only sources in Java; Bedrock adds raid drops. The same holds for Frost Walker, Curse of Binding and Curse of Vanishing among the books a librarian sells.

Does curing a zombie villager make Mending cost 1 emerald?

Not always. A cure takes 20 emeralds off permanently, and a Mending book lists at 10 to 38, so the cured price runs from 1 to 18 emeralds and averages 6.4. Six of the fifteen prices a Mending book can list are 20 emeralds or less, and only those drop to the one-emerald floor — 40% of the time. The rest of the time you pay between 2 and 18.

Why did my librarian stop changing trades?

You traded with it. One completed trade locks a villager's offers permanently, and breaking the lectern afterwards only removes the workstation. Reroll before you trade anything, including the 24-paper trade at the novice tier.

Is it faster to trade for Fortune III or enchant it at a table?

The table, by a wide margin. Fortune III averages about 5 tries at 15 bookshelves on the best slot, at 3 levels and 3 lapis a try, against 180 lectern resets at a librarian. The same holds for Unbreaking III at 2 tries, Protection IV at 4 and Looting III at 10. Trading wins on the four treasure books and on level windows a table cannot reach, such as Efficiency V on a diamond pickaxe. Compare your own item and bookshelf count with the Minecraft Enchanting Probability Calculator.

Why is Efficiency V almost always 64 emeralds?

The game rolls a level V book's price as an integer from 17 to 71, then clips anything above 64 down to 64. Seven of the 55 possible rolls sit above 64 and are pulled down to it, so eight rolls in all end at 64 — eight times more often than any other price. Waiting for a cheap one is expensive: Efficiency V alone averages 300 resets, and holding out for the 17-emerald listing pushes it to 16,500.

Do Bedrock librarians have worse odds than Java?

Yes, by a third. A Java librarian unlocks two trades per level from that level's pool, so 2 in 3 fresh novices offer an enchanted book. A Bedrock librarian fills each slot one trade at a time and the book shares its slot with one other trade, giving 1 in 2. Mending goes from 1 in 60 to 1 in 80, and every reset count grows by the same 4/3.

How many librarians do I need for every enchantment?

257 fresh Java librarians on average to see all 40 enchantments at least once at any level, 933 to see every one at its highest level, and 8,695 if each also has to be cheap enough that a cured villager sells it for a single emerald. On Bedrock those three figures are 342, 1,244 and 11,594. The tail dominates: the last few enchantments take longer to appear than the first thirty put together.

Does leveling the librarian give me another shot at the book?

It gives a new chance, never a better one. A Java librarian taken to Master has had about 2.5 book trades rolled across its levels, but only the first can be rerolled with the lectern, and trading once locks the whole list. No amount of leveling upgrades an Efficiency I already on the list into an Efficiency V.

Do I need my world seed to use this calculator?

No. It works from the game's published trade tables, so it gives the odds across every world rather than the sequence in yours. Java 26.1 did tie the reroll sequence to the seed, so a seed-aware predictor can name your next trade — a different question, and a good complement to this one. Nothing you type here leaves your browser.

Is this librarian calculator free, and how accurate is it?

It is free and needs no account. Every constant behind it — the 40-enchantment tradeable pool, the per-level trade pools for Java and Bedrock, the price window running from 2 + 3 × level to 6 + 13 × level with its treasure doubling and 64-emerald ceiling, and the reputation values a cure moves — is read from the vanilla data pack at Minecraft 26.2 (the 1.21 line, renamed) and cross-checked against the Minecraft Wiki. The reset counts are exact expected values — the reciprocal of the per-reset probability, with no sampling involved.

The calculator says 300 resets. Should I keep going?

Weigh three numbers before deciding. First the table verdict: if an enchanting table can reach the book, it wins on effort almost every time. Second the nine-in-ten figure, because an average of 300 means one hunt in ten runs past 690. Third the time card at your own seconds per reset, because 300 resets at 8 seconds is 40 minutes of breaking a lectern. If all three still favor the librarian, raising your emerald cap is the cheapest way to shorten the hunt.

Sources & References

Content verified by the Smart Calculators Team