Skip the island and read the website

Hash Generator — SHA-1, SHA-256, SHA-384, SHA-512

← Safe HarborThe Creative Island · Island 14

SHA-256 and friends, computed by your browser's own crypto engine — text in, digest out, nothing uploaded.

One-way by design: the same text always gives the same digest, but the digest can never be reversed to the text.

How to use this tool

  1. Type or paste your text (hashing starts automatically).
  2. Read the four digests — SHA-256 is today's general-purpose standard.
  3. Copy the SHA-256 digest if you need it for verification or deduplication.

Example

hello → SHA-256 2cf24dba5fb0a30e26e83b2ac5b9e29e1b161e5c1fa7425e73043362938b9824. Change one letter — 'Hello' — and the digest becomes completely different: 185f8db32271fe25f561a6fc938b2e26… This is the avalanche effect.

The formula

digest = SHA(text) — fixed-length output (160/256/384/512 bits) regardless of input size

What the result means

A hash is a fingerprint: deterministic (same input, same digest), avalanche (any change rewrites everything), and one-way (no algorithm recovers the input). That trio powers file integrity checks, password storage (with salting), deduplication and blockchain links. SHA-1 is shown only because legacy systems still ask for it — for anything new, use SHA-256 or better.

Frequently asked questions

Why is SHA-1 in the list if it's broken?

Collision attacks exist against SHA-1, so it's retired for security uses — but terabytes of legacy systems still request it. It's here for compatibility, marked with a warning label in spirit.

Can a hash be reversed to reveal the original text?

No — not by computation. 'Reversal' sites just look up precomputed tables of common inputs; unique, random inputs have no entries to find.

How are passwords hashed then?

With special password algorithms (bcrypt, Argon2) that are deliberately slow and salted — SHA alone is too fast, making leaked-hash cracking cheaper. Different tools for different jobs.

Is my text uploaded?

No — hashing happens in your browser via the Web Crypto API.

More tools from Safe Harbor

Explore the island world

Random Token Generator — API Keys & Secrets

← Safe HarborThe Creative Island · Island 14

API keys, hex secrets, URL-safe tokens — generated with cryptographic randomness, on your device.

Generate client-side only for local use or your own system. Real production API keys must come from the service that will verify them — never from a web page.

How to use this tool

  1. Choose the byte length (32 bytes / 256 bits is the modern standard for secrets).
  2. Pick a format: hex for classic API keys, URL-safe for paths and query strings, Base64 for compact tokens.
  3. Generate, copy, and store it somewhere safe — there is no 'recover' button for random secrets.

Example

A 32-byte hex token looks like 9f86d081884c7d659a2feaa0c55ad015… — 256 bits of entropy, meaning every possible key is equally likely and guessing is hopeless at any scale.

The formula

crypto.getRandomValues → N random bytes → hex (2 chars/byte), URL-safe base64url, or 6-bit alphabet

What the result means

Tokens differ from passwords in one way: no human ever types them, so length and uniform randomness are the only design constraints — and both are free. 256 bits is the standard because it leaves brute-force dead for any future hardware we can imagine, and because hashing (which halves the search space) still leaves it uncrackable.

Frequently asked questions

Can I use these as real API keys?

For local projects, testing and your own systems, yes. For a production service, the issuing server must generate the key (it stores a hash to verify) — a page-generated string can't be verified by anyone but you.

What length should I pick?

32 bytes (256 bits) for anything. 16 bytes (128 bits) is acceptable for short-lived or low-value tokens.

Why URL-safe instead of standard Base64?

Standard Base64 contains + and /, which break URLs and filenames; the URL-safe alphabet swaps them for - and _ and drops padding.

Are generated tokens stored anywhere?

No — generation is local; refresh the page and it is gone forever.

More tools from Safe Harbor

Explore the island world

Random Group Picker — Fair Teams & Winners

← Safe HarborThe Creative Island · Island 14

Paste the names, pick how many groups — get fair, unbiased teams in one click.

How to use this tool

  1. Paste your names, one per line.
  2. Choose the number of groups (or switch to winner-picking mode).
  3. Press Shuffle — unbiased groups, balanced in size.

Example

24 students into 4 groups: three shuffled teams of 6, no clustering of friends, no 'teacher's pets together' accusations — the loudest classroom argument, deleted in one click.

The formula

Fisher-Yates shuffle driven by crypto randomness → every arrangement equally likely

What the result means

Random grouping kills two biases at once: the organizer's unconscious bias (friends, strong-with-strong) and the group's suspicion of it. Fisher-Yates with a cryptographic source is the gold standard — every permutation has exactly equal probability, which no 'names in a hat' with awkward folding can promise.

Frequently asked questions

Are the groups really random?

Yes — Fisher-Yates shuffle on crypto randomness: every possible grouping is equally likely, with no pattern a participant could predict.

Can I pick a single winner?

Yes — switch mode to 'Pick N winners only' and set the count.

Why is one group smaller?

When names don't divide evenly, the first groups get the extra member — sizes never differ by more than one.

Are the names uploaded?

No — the shuffle happens in your browser; the name list never leaves your device.

More tools from Safe Harbor

Explore the island world

Random Number Generator — Fair Draws, Any Range

← Safe HarborThe Creative Island · Island 14

Pick a number between anything and anything — once or many times, with or without repeats.

How to use this tool

  1. Set your range (1–100, 1–6 for dice, 1–49 for lottery style).
  2. Choose how many numbers and whether repeats are allowed.
  3. Press Draw — each press is a fresh, independent draw.

Example

Class raffle with 60 students: From 1, To 60, How many 3, No repeats → three distinct winners in one click, immune to the 'whose handwriting wins' effect of paper draws.

The formula

rejection-sampled crypto.getRandomValues → uniform integers, no modulo bias

What the result means

True fairness in draws comes from uniform randomness — this generator rejects out-of-range values before mapping (avoiding modulo bias) and uses the browser's cryptographic source, the same one that guards HTTPS. That's a stronger guarantee than dice, and far stronger than 'the boss picks'.

Frequently asked questions

Can I use it for a raffle or giveaway?

Yes — it's a fair draw: every number has identical probability. For public trust, screenshot the result.

Why no repeats by default?

Most draws (winners, orderings) want unique picks. Uncheck it for independent events like dice rolls.

Is the randomness really cryptographic?

Yes — crypto.getRandomValues, seeded by your operating system's secure entropy pool. Math.random is only used as a fallback on ancient browsers.

Are my draws recorded?

No — nothing is stored or sent; refresh and the history is gone.

More tools from Safe Harbor

Explore the island world

Password Strength Checker — Test Without Judgement

← Safe HarborThe Creative Island · Island 14

Type a test password — see its entropy, estimated crack time, and every weakness. Stays on your device, guaranteed.

Test with a similar-shaped password if you are cautious — same length and pattern, different characters. The analysis works the same.

How to use this tool

  1. Type or paste the password to test — it stays in this page's memory only (use the show/hide toggle).
  2. Read the entropy bits, the verdict bar and the crack-time estimate.
  3. Fix every finding the checker lists, then re-test.

Example

P@ssword123 tests as 10 chars but the checker finds the common word, the word+digits pattern and the substitution — effective strength collapses to WEAK. A random 16-character output from the Password Forge scores ~99 bits: VERY STRONG.

The formula

effective bits = effective length (patterns removed) × log₂(character pool) · crack time at 10¹² guesses/sec

What the result means

Crackers don't brute-force blindly — they try wordlists, patterns and substitution rules first, so 'P@ssword1' falls in seconds despite its 9 characters. Real strength = randomness the attacker cannot model. The crack-time figure assumes an offline attack on a leaked hash, the worst realistic case.

Frequently asked questions

Is it safe to paste my real password here?

The page has no network calls and no storage — the check runs in your browser only. If uneasy, test a similar-shaped dummy (same length, same pattern).

What is a good target?

70+ bits for accounts that matter, 100+ for master passwords and encryption keys. Length is the cheapest way up: each random character adds 5-6 bits.

Why does the checker say 8 random chars are only FAIR?

Because GPU rigs really do crack random 8-character hashes in hours when they steal the hash. Online logins are rate-limited; leaked-password cracking is not.

Does adding '2024' to a word help?

It adds a pattern, not strength — 'word+year' is among the first rules every cracking engine tries.

More tools from Safe Harbor

Explore the island world

Password Generator — Strong Random Passwords, Locally

← Safe HarborThe Creative Island · Island 14

Cryptographically random passwords — your length, your character sets, generated on your own device.

How to use this tool

  1. Set your length — 16+ is the modern default; 20+ for anything precious.
  2. Tick the character sets (the site requiring the password will state its rules).
  3. Press Generate and Copy — then store it in a password manager, never a text file.

Example

A 16-character password from all four sets carries ~99 bits of entropy — offline brute-force at a trillion guesses per second would outlive the planet. The same 16 characters from lowercase-only carries just 75 bits, and 'P@ssword1' carries about 10.

The formula

entropy = length × log₂(pool size) · randomness from crypto.getRandomValues (Web Crypto)

What the result means

Password strength is almost entirely a function of randomness and length — substituting '@' for 'a' adds nothing if a human chose it, because crackers try those substitutions first. True random output has no patterns to model, which is why generated passwords can be shorter than memorable ones and still stronger.

Frequently asked questions

Can this site see or store my password?

No. Generation happens in your browser with your device's own random generator; the password is displayed locally and never transmitted. Closing the tab erases it forever.

Is 8 characters still okay?

Not for anything you care about. Modern GPU rigs crack random 8-character passwords in hours. Use 16+ generated characters, or long passphrases where humans must remember.

Why avoid lookalike characters (I, l, 1, O, 0)?

The generator quietly omits them — they cause transcription errors when a password must be typed from paper or read over the phone.

How am I supposed to remember these?

You are not — that is the password manager's job. Remember one strong master passphrase; let the manager hold the rest.

More tools from Safe Harbor

Explore the island world

Probability Calculator — Percent, Odds & One-in-N

← Research ReefThe Creative Island · Island 13

What are the chances? Fractions to percentages, percentages to one-in-N, odds made honest.

Dice quick facts: any single face 1/6 ≈ 16.7% · double sixes 1/36 ≈ 2.8% · coin flip 50% · at least one six in four rolls ≈ 51.8%.

How to use this tool

  1. Enter the favorable outcomes and the total outcomes (e.g. 1 and 6 for one die face).
  2. Press Calculate probability — get the percentage, fraction, one-in-N and odds.
  3. The percentage shortcut converts a known percentage the other way.

Example

One specific face of a die: 1/6 = 16.67% = one in 6. Rolling at least one six in four rolls: 1 − (5/6)&sup4; ≈ 51.8% — slightly better than a coin flip, a result most people guess wrong.

The formula

P = favorable ÷ total  |  odds against = (1−P) : P  |  'at least one' = 1 − (1−p)™

What the result means

Probability is long-run frequency, not destiny: a one-in-six event can happen three times in a row, and the die owes you nothing afterwards. The one-in-N form builds intuition ('one in 6 million' lands harder than '0.0000167%'), and 'at least one' problems hide inside exams, lotteries and quality control everywhere.

Frequently asked questions

Why is 'at least one' calculated with subtraction?

Because the direct cases pile up awkwardly. The complement — the event NOT happening every time — is one multiplication: 1 − (5/6)&sup4; for four dice. This trick unlocks a huge family of problems.

Do independent events even out after a streak?

No — the gambler's fallacy. Past rolls don't change future odds; the die has no memory and owes no debts.

Can probability be over 100%?

Never. If your arithmetic says so, some outcomes were double-counted — the classic P(A)+P(B) error when events overlap.

Is my input stored?

No — the calculation happens in your browser.

More tools from Research Reef

Explore the island world

Fish Stocking Calculator — How Many Fish Fit Your Tank

← Research ReefThe Creative Island · Island 13

The eternal question — how many fish? Volume, fish size, and a rule of thumb with its limits stated honestly.

Rule-of-thumb stocking is a starting point, not a law — research your species' adult size, shoaling needs, territory and bioload before buying.

How to use this tool

  1. Enter your tank's actual water volume in litres (use the Aquarium Volume Calculator for the honest figure).
  2. Enter the fish's adult size in cm — not the size in the shop tank.
  3. Choose a stocking approach and press Calculate stocking.

Example

A 56 L tank with 5 cm adult tetras, standard rule: 56 ÷ 5 ≈ 11 fish — a proper shoal. The same tank with 12 cm goldfish: 4 fish by arithmetic, but honestly fewer, because goldfish are messy bioload machines.

The formula

fish count ≈ volume ÷ (adult cm × litres-per-cm factor)

What the result means

The litres-per-fish rule encodes a biological truth: water volume and filtration set how much waste bacteria can convert. It ignores species behavior — territory, shoaling, aggression — which is why the result is a ceiling to think under, not a target to hit. Adult size matters most: the 3 cm shop juvenile that grows to 15 cm turns a stocked tank into an overcrowded one.

Frequently asked questions

Is '1 cm per litre' actually reliable?

It's a useful sanity ceiling for small peaceful community fish, and wrong for big, messy or territorial species. Use it to reject bad ideas ('20 goldfish in 40L'), not to design a biotope.

Should I stock to the maximum?

No — experienced keepers stock 50–70% of the ceiling. Understocked tanks are stable, forgiving and easier to keep beautiful.

Does it work for bettas and other solitary fish?

Volume-wise yes, but solitary/territorial species override the count: one betta regardless of what the arithmetic allows, and territorial cichlids need footroom, not just litres.

Is my aquarium data uploaded?

No — the estimate is computed in your browser.

More tools from Research Reef

Explore the island world

Aquarium Volume Calculator — Litres & Gallons of Your Tank

← Research ReefThe Creative Island · Island 13

How much water is really in there? Rectangular and cylinder tanks, litres and gallons, honest about the air gap at the top.

How to use this tool

  1. Choose your tank shape: rectangular or cylinder.
  2. Measure in cm: length × width × height (or diameter × height).
  3. Set the fill level — water always sits below the rim — and read litres and gallons.

Example

A standard 60 × 30 × 35 cm tank: gross 63 L, at 90% fill ≈ 56.7 L — and with a gravel bed and wood, realistically a ~48 L working volume. That's the number your stocking and dosing decisions should use.

The formula

rectangular: L×W×H ÷ 1000 = litres  |  cylinder: π×(D÷2)²×H ÷ 1000

What the result means

Two mistakes ruin tank math: quoting the gross volume (nobody fills to the rim) and ignoring displacement (gravel, wood and filter media replace water). Dosing medication or fertilizer by the gross figure overdoses a smaller real volume — this calculator nudges you toward the honest number.

Frequently asked questions

Why does the tank shop say 60L but this says 56?

Shops quote gross (or marketing) volume. Real water level, glass thickness and decor take 10–20% away — always dose by actual litres.

How do I account for a sump or canister filter?

Add the filter's water volume (sump water depth × area, or ~2–4 L for a typical canister) to the tank's figure for total system volume.

Does shape affect stocking?

Yes — surface area matters more than litres for gas exchange; a long shallow 60L outperforms a tall narrow 60L. Volume is necessary math, not the whole story.

Is my tank data uploaded?

No — computed locally in your browser.

More tools from Research Reef

Explore the island world