Filter
Defines composable checks that can also transform values.
A Filter<Input, Pass, Fail> receives an input and returns a Result.
Success means the value passed the filter, and failure means the value was
filtered out. Filters may also narrow or transform the passing value. This
module includes constructors from predicates, options, and effects, built-in
filters for common JavaScript values and tags, helpers for combining filters,
and conversions to predicates, options, and results.
Combinators
Combines two filters but only returns the result of the left filter.
Signature
declare const andLeft: { <InputR, PassR, FailR>(right: Filter<InputR, PassR, FailR>): <InputL, PassL, FailL>(left: Filter<InputL, PassL, FailL>) => Filter<InputL & InputR, PassL, FailR | FailL>; <InputL, PassL, FailL, InputR, PassR, FailR>(left: Filter<InputL, PassL, FailL>, right: Filter<InputR, PassR, FailR>): Filter<InputL & InputR, PassL, FailL | FailR>;}Example
(Keeping the left filter result)
import { Filter, Result } from "effect"
const positiveNumbers = Filter.fromPredicate((n: number) => n > 0)const evenNumbers = Filter.fromPredicate((n: number) => n % 2 === 0)
const positiveEven = Filter.andLeft(positiveNumbers, evenNumbers)positiveEven(2) // => Result.succeed(2)Combines two filters but only returns the result of the right filter.
Signature
declare const andRight: { <InputR, PassR, FailR>(right: Filter<InputR, PassR, FailR>): <InputL, PassL, FailL>(left: Filter<InputL, PassL, FailL>) => Filter<InputL & InputR, PassR, FailR | FailL>; <InputL, PassL, FailL, InputR, PassR, FailR>(left: Filter<InputL, PassL, FailL>, right: Filter<InputR, PassR, FailR>): Filter<InputL & InputR, PassR, FailL | FailR>;}Example
(Keeping the right filter result)
import { Filter, Result } from "effect"
const positiveNumbers = Filter.fromPredicate((n: number) => n > 0)const doubleNumbers = Filter.make((n: number) => n > 0 ? Result.succeed(n * 2) : Result.fail(n))
const positiveDoubled = Filter.andRight(positiveNumbers, doubleNumbers)positiveDoubled(2) // => Result.succeed(4)Composes two filters sequentially, feeding the output of the first into the second.
Signature
declare const compose: { <PassL, PassR, FailR>(right: Filter<PassL, PassR, FailR>): <InputL, FailL>(left: Filter<InputL, PassL, FailL>) => Filter<InputL, PassR, FailR | FailL>; <InputL, PassL, FailL, PassR, FailR>(left: Filter<InputL, PassL, FailL>, right: Filter<PassL, PassR, FailR>): Filter<InputL, PassR, FailL | FailR>;}Example
(Composing filters)
import { Filter, Result } from "effect"
const stringFilter = Filter.stringconst nonEmptyUpper = Filter.make((s: string) => s.length > 0 ? Result.succeed(s.toUpperCase()) : Result.fail(s))
const stringToUpper = Filter.compose(stringFilter, nonEmptyUpper)stringToUpper("hello") // => Result.succeed("HELLO")composePassthrough
Composes two filters sequentially, passing the successful output of the first filter to the second.
Details
If either filter fails, the returned filter fails with the original input instead of the intermediate failure value.
Signature
declare const composePassthrough: { <InputL, PassL, PassR, FailR>(right: Filter<PassL, PassR, FailR>): <FailL>(left: Filter<InputL, PassL, FailL>) => Filter<InputL, PassR, InputL>; <InputL, PassL, FailL, PassR, FailR>(left: Filter<InputL, PassL, FailL>, right: Filter<PassL, PassR, FailR>): Filter<InputL, PassR, InputL>;}Combines two filters with logical OR semantics.
Signature
declare const or: { <Input2, Pass2, Fail2>(that: Filter<Input2, Pass2, Fail2>): <Input1, Pass2, Fail2>(self: Filter<Input1, Pass2>) => Filter<Input1 & Input2, Pass2, Fail2>; <Input1, Pass1, Fail1, Input2, Pass2, Fail2>(self: Filter<Input1, Pass1, Fail1>, that: Filter<Input2, Pass2, Fail2>): Filter<Input1 & Input2, Pass1 | Pass2, Fail2>;}Combines two filters into a tuple of their results.
Details
Both filters must succeed for the combination to succeed. If both pass, their outputs are combined into a tuple.
Signature
declare const zip: { <InputR, PassR, FailR>(right: Filter<InputR, PassR, FailR>): <InputL, PassL, FailL>(left: Filter<InputL, PassL, FailL>) => Filter<InputL & InputR, [PassL, PassR], FailR | FailL>; <InputL, PassL, FailL, InputR, PassR, FailR>(left: Filter<InputL, PassL, FailL>, right: Filter<InputR, PassR, FailR>): Filter<InputL & InputR, [PassL, PassR], FailL | FailR>;}Example
(Zipping filters)
import { Filter, Result } from "effect"
const positiveNumbers = Filter.fromPredicate((n: number) => n > 0)const evenNumbers = Filter.fromPredicate((n: number) => n % 2 === 0)
const positiveAndEven = Filter.zip(positiveNumbers, evenNumbers)positiveAndEven(2) // => Result.succeed([2, 2])Combines two filters and applies a function to their results.
When to use
Use to combine two filters with a custom function to merge their outputs.
Details
Both filters must succeed (not return fail) for the combination to succeed.
If both filters pass, their outputs are combined using the provided function.
See
- zip for combining two filters into a tuple
Signature
declare const zipWith: { <PassL, InputR, PassR, FailR, A>(right: Filter<InputR, PassR, FailR>, f: (left: PassL, right: PassR) => A): <InputL, FailL>(left: Filter<InputL, PassL, FailL>) => Filter<InputL & InputR, A, FailR | FailL>; <InputL, PassL, FailL, InputR, PassR, FailR, A>(left: Filter<InputL, PassL, FailL>, right: Filter<InputR, PassR, FailR>, f: (left: PassL, right: PassR) => A): Filter<InputL & InputR, A, FailL | FailR>;}Constructors
A predefined filter that only passes through bigint primitive values.
When to use
Use to keep primitive big integer values from unknown input while staying in
the composable Filter / Result pipeline.
Details
Implemented with fromPredicate(Predicate.isBigInt), so values where
typeof input === "bigint" succeed and all other inputs fail with the
original input.
Gotchas
This filter does not coerce numbers or strings; 1n passes while 1 fails.
See
- number for JavaScript
numbervalues - Predicate.isBigInt for the underlying guard
Signature
declare const bigint: Filter<unknown, bigint>A predefined filter that only passes through boolean values.
When to use
Use when accepting an unknown input only if it is already a boolean and you
want a Filter result rather than a plain predicate result.
Details
Implemented with fromPredicate(Predicate.isBoolean), so true and false
succeed and non-booleans fail with the original input.
See
- Predicate.isBoolean for the underlying guard
- fromPredicate for custom predicate-based filters
Signature
declare const boolean: Filter<unknown, boolean>A predefined filter that only passes through Date objects.
When to use
Use when you need to narrow unknown input to JavaScript Date instances with
a reusable Filter.
Details
Implemented with fromPredicate(Predicate.isDate), so passing values return
Result.succeed(input) and failing values return Result.fail(input).
Gotchas
The check uses instanceof Date, so invalid Date objects still pass; the
filter does not validate the timestamp.
See
- Predicate.isDate for the underlying guard
- instanceOf for constructor-based filtering
- fromPredicate for custom date checks
Signature
declare const date: Filter<unknown, Date>Creates a filter that only passes values equal to the specified value using structural equality.
When to use
Use to accept inputs that are structurally equal to a known expected value
while staying in a composable Filter / Result pipeline.
Details
Delegates to Equal.equals. On success it returns Result.succeed(value);
on failure it returns Result.fail(input).
See
- equalsStrict for JavaScript
===matching instead of structural equality - Equal.equals for the underlying structural equality semantics
Signature
declare function equals<A, Input = unknown>(value: A): Filter<Input, A, EqualsWith<Input, A, A, Exclude<Input, A>>>equalsStrict
Creates a Filter that passes only values strictly equal to the specified
value using JavaScript === comparison.
When to use
Use when you need a Filter that accepts only the exact primitive value or
object reference using JavaScript strict equality in a Filter / Result
pipeline.
Gotchas
NaN never passes, even when the expected value is NaN, and objects pass
only when they are the same reference.
See
- equals for structural equality when distinct values with equal contents should pass
Signature
declare function equalsStrict<A, Input = unknown>(value: A): Filter<Input, A, EqualsWith<Input, A, A, Exclude<Input, A>>>fromPredicate
Creates a Filter from a predicate or refinement function.
Details
This is a convenient way to create filters from boolean-returning functions.
When the predicate returns true, the input value is passed through unchanged.
When it returns false, the fail type is returned.
Signature
declare const fromPredicate: { <A, B>(refinement: Refinement<A, B>): Filter<A, B, EqualsWith<A, B, A, Exclude<A, B>>>; <A>(predicate: Predicate<A>): Filter<A>;}Example
(Creating filters from predicates)
import { Filter, Result } from "effect"
// Create filter from predicateconst positiveNumbers = Filter.fromPredicate((n: number) => n > 0)const nonEmptyStrings = Filter.fromPredicate((s: string) => s.length > 0)
// Type refinementconst isString = Filter.fromPredicate((x: unknown): x is string => typeof x === "string")positiveNumbers(1) // => Result.succeed(1)nonEmptyStrings("") // => Result.fail("")isString("ok") // => Result.succeed("ok")fromPredicateOption
Creates a Filter from a function that returns an Option; Some(value)
passes with value, and None fails with the original input.
Signature
declare function fromPredicateOption<A, B>(predicate: (a: A) => Option<B>): Filter<A, B>Creates a Filter that passes inputs whose has(key) method returns
true for the specified key.
When to use
Use to keep inputs that expose a has method, such as Set or Map, when
they contain a required key.
See
- fromPredicate for custom predicate filters or inputs without a
hasmethod - Predicate.hasProperty for guarding property presence instead of
calling an input's
hasmethod
Signature
declare function has<K>(key: K): <Input extends { readonly has: (key: K) => boolean;}>(input: Input) => Result<Input, Input>instanceOf
Creates a filter that only passes instances of the given constructor.
When to use
Use to narrow unknown input to values created by a specific JavaScript
constructor while keeping the result in the Filter / Result pipeline.
Details
The filter succeeds when the input satisfies instanceof constructor.
Otherwise it fails with the original input.
Gotchas
This uses JavaScript instanceof semantics, including prototype-chain and
realm behavior.
See
- fromPredicate for custom predicate-based narrowing
Signature
declare function instanceOf<K extends (...args: any) => any>(constructor: K): <Input>(u: Input) => Result<InstanceType<K>, Exclude<Input, InstanceType<K>>>Creates a Filter from a function that returns either a pass or fail value.
Details
This is the primary constructor for creating custom filters. The function
should return either Result.succeed(value) or Result.fail(value).
Signature
declare function make<Input, Pass, Fail>(f: (input: Input) => Result<Pass, Fail>): Filter<Input, Pass, Fail>Example
(Creating custom filters)
import { Filter, Result } from "effect"
// Create a filter for positive numbersconst positiveFilter = Filter.make((n: number) => n > 0 ? Result.succeed(n) : Result.fail(n))
// Create a filter that transforms strings to uppercaseconst uppercaseFilter = Filter.make((s: string) => s.length > 0 ? Result.succeed(s.toUpperCase()) : Result.fail(s))positiveFilter(1) // => Result.succeed(1)uppercaseFilter("ok") // => Result.succeed("OK")makeEffect
Creates an effectful Filter from a function that returns an Effect.
Details
This constructor is used when the filtering operation needs to perform effectful computations, such as async operations, error handling, or accessing services from the environment.
Signature
declare function makeEffect<Input, Pass, Fail, E, R>(f: (input: Input) => Effect<Result<Pass, Fail>, E, R>): FilterEffect<Input, Pass, Fail, E, R>Example
(Creating effectful filters)
import { Effect, Filter, Result } from "effect"
// Create an effectful filter that validates asyncconst asyncValidate = Filter.makeEffect((id: string) => Effect.gen(function*() { const isValid = yield* Effect.succeed(id.length > 0) return isValid ? Result.succeed(id) : Result.fail(id) }))
await Effect.runPromise(asyncValidate("id")) // => Result.succeed("id")A predefined filter that only passes through number values.
Signature
declare const number: Filter<unknown, number>Example
(Filtering numbers)
import { Filter, Result } from "effect"
Filter.number(42) // => Result.succeed(42)Filter.number("42") // => Result.fail("42")Creates a filter that extracts a reason from a tagged error.
Signature
declare const reason: { <Input>(): <Tag extends string, ReasonTag extends string>(tag: Tag, reasonTag: ReasonTag) => Filter<Input, ExtractReason<ExtractTag<Input, Tag>, ReasonTag>, Input>; <Input, Tag extends string, ReasonTag extends string>(tag: Tag, reasonTag: ReasonTag): Filter<Input, ExtractReason<ExtractTag<Input, Tag>, ReasonTag>, Input>; <Tag extends string, ReasonTag extends string>(tag: Tag, reasonTag: ReasonTag): <Input>(input: Input) => Result<ExtractReason<ExtractTag<Input, Tag>, ReasonTag>, Input>;}A predefined filter that only passes through string values.
Signature
declare const string: Filter<unknown, string>Example
(Filtering strings)
import { Filter, Result } from "effect"
Filter.string("hello") // => Result.succeed("hello")Filter.string(42) // => Result.fail(42)A predefined filter that only passes through Symbol values.
Signature
declare const symbol: Filter<unknown, symbol>Creates a filter that checks if an input is tagged with a specific tag.
When to use
Use to keep only the matching member of a _tag-discriminated union while
staying in a composable Filter / Result pipeline.
Details
The filter succeeds when Predicate.isTagged(input, tag) returns true.
Otherwise it fails with the original input.
Gotchas
This only checks _tag; it does not validate the rest of the variant fields.
See
- Predicate.isTagged for the underlying boolean guard when a
Filterresult is not needed - reason for extracting a nested reason variant from tagged errors
Signature
declare const tagged: { <Input>(): <Tag extends string>(tag: Tag) => Filter<Input, ExtractTag<Input, Tag>, ExcludeTag<Input, Tag>>; <Input, Tag extends string>(tag: Tag): Filter<Input, ExtractTag<Input, Tag>, ExcludeTag<Input, Tag>>; <Tag extends string>(tag: Tag): <Input>(input: Input) => Result<ExtractTag<Input, Tag>, ExcludeTag<Input, Tag>>;}Converting
Converts a Filter into a function that returns Some for passed values
and None for filtered-out values.
When to use
Use when adapting a Filter to Option-based code where passed values
become Some and filtered-out inputs become None.
See
- toResult for keeping the filter failure value
- toPredicate for plain boolean pass/fail checks
Signature
declare function toOption<A, Pass, Fail>(self: Filter<A, Pass, Fail>): (input: A) => Option<Pass>toPredicate
Converts a Filter into a predicate function.
When to use
Use to reuse a Filter with APIs that accept only boolean predicates when
the pass and fail payloads are not needed.
See
Signature
declare function toPredicate<A, Pass, Fail>(self: Filter<A, Pass, Fail>): Predicate<A>Converts a Filter into a function that returns the underlying
Result.Result for each input.
When to use
Use to adapt a Filter to APIs that expect a plain function returning
Result, while preserving both the pass value and the failure value.
See
- toOption for keeping only passed values
- toPredicate for plain boolean pass/fail checks
Signature
declare function toResult<A, Pass, Fail>(self: Filter<A, Pass, Fail>): (input: A) => Result<Pass, Fail>Mapping
Transforms the failure value produced by a Filter, leaving successful
results unchanged.
Signature
declare const mapFail: { <Fail, Fail2>(f: (fail: Fail) => Fail2): <Input, Pass>(self: Filter<Input, Pass, Fail>) => Filter<Input, Pass, Fail2>; <Input, Pass, Fail, Fail2>(self: Filter<Input, Pass, Fail>, f: (fail: Fail) => Fail2): Filter<Input, Pass, Fail2>;}Models
Represents a filter function that can transform inputs to outputs or filter them out.
Details
A filter takes an input value and either returns a boxed pass value or the
special fail type to indicate the value should be filtered out.
Signature
interface Filter<in Input, out Pass = Input, out Fail = Input> { (input: Input): Result<Pass, Fail>;}Example
(Defining a positive number filter)
import { Filter, Result } from "effect"
// A filter that only passes positive numbersconst positiveFilter: Filter.Filter<number> = (n) => n > 0 ? Result.succeed(n) : Result.fail(n)
positiveFilter(5) // => Result.succeed(5)positiveFilter(-3) // => Result.fail(-3)FilterEffect interface
Represents an effectful filter function that can produce Effects.
Details
Similar to a regular Filter, but the filtering operation itself can be
effectful, allowing for asynchronous operations, error handling, and
dependency injection.
Signature
interface FilterEffect<in Input, out Pass, out Fail, out E = never, out R = never> { (input: Input): Effect<Result<Pass, Fail>, E, R>;}Example
(Defining an effectful user filter)
import { Effect, Filter, Result } from "effect"
// An effectful filter that validates user datatype User = { id: string; isActive: boolean }type ValidationError = { message: string }
const validateUser: Filter.FilterEffect< string, User, User, ValidationError, never> = (id) => Effect.gen(function*() { const user: User = { id, isActive: id.length > 0 } return user.isActive ? Result.succeed(user) : Result.fail(user) })
await Effect.runPromise(validateUser("alice")) // => Result.succeed({ id: "alice", isActive: true })await Effect.runPromise(validateUser("")) // => Result.fail({ id: "", isActive: false })