Many Solana accounts are controlled by a wallet keypair, but programs also need predictable accounts for settings, user records, and token vaults. A Program Derived Address, or PDA, gives a program a repeatable address without creating another private key for someone to manage.
What is a PDA?
A PDA is a 32-byte address derived from a program ID and a set of seeds. Seeds are inputs chosen by the program, such as a fixed label, a user address, or a token mint. If the same program uses the same seeds, it derives the same address every time. Changing the program ID or a seed gives a different result.
The important detail is that a PDA is deliberately off the Ed25519 curve used by ordinary Solana public keys. That means there is no corresponding private key that a person can hold. The program that derived the PDA can authorize it during a cross-program invocation by supplying the correct seeds to the runtime through invoke_signed. This is program authorization, not a hidden wallet signature.
How seeds and bumps work
The SDK hashes the seeds together with the program ID and a fixed marker, then checks whether the result is off-curve. If a candidate lands on the curve, it tries a one-byte value called a bump. The standard search begins at 255 and moves downward until it finds a valid off-curve address. That first valid value is the canonical bump, and programs should consistently use it so one seed recipe maps to one expected address.
Solana limits a derivation to at most 16 seeds, with each seed no longer than 32 bytes. Those limits matter when an application turns user input into seeds: the program should define a clear format and validate every input before using it.
A PDA is not automatically an account
Calculating a PDA only calculates an address. It does not create storage at that address, put SOL there, or initialize account data. A program must explicitly create and initialize an account at the PDA, often by invoking the System Program. The derived address and the on-chain account are related, but they are different things.
Why applications use PDAs
PDAs are useful for records that should be easy to find from known inputs: one profile per user, one configuration account for a program, or a vault associated with a user and a token. A client can calculate the expected address in advance, while the program retains the authority to sign for it. This avoids storing a separate secret key for every program-managed account.
Safety checks for users and developers
- Check which program ID derives and controls the PDA; a familiar-looking address alone is not proof of trust.
- Use the canonical bump and validate that supplied accounts match the expected seeds.
- Keep seed boundaries unambiguous. Since seeds are concatenated before hashing, poorly designed variable-length inputs can create confusing overlaps.
- Remember that a program can control only PDAs derived under its own program ID; another program cannot simply claim the same authority.
A PDA is best understood as a predictable, program-controlled address—not as a wallet with a missing password.
When reviewing a Solana app, ask what seeds define an account, which program owns the derivation, and whether the address is actually initialized. Those checks make PDA-based accounts easier to understand and help users distinguish deterministic program state from ordinary wallet funds.