<A, E, R>(effect: Effect.Effect<A, E, R>): Channel<
A,
Pull.ExcludeDone<E>,
void,
unknown,
unknown,
unknown,
R
>Uses an effect to write a single value to the channel.
Example (Creating channels from effects)
import { Channel, Data, Effect } from "effect"
class DatabaseError extends Data.TaggedError("DatabaseError")<{
readonly message: string
}> {}
// Create a channel from a successful effect
const successChannel = Channel.fromEffect(
Effect.succeed("Hello from effect!")
)
// Create a channel from an effect that might fail
const fetchUserChannel = Channel.fromEffect(
Effect.tryPromise({
try: () => fetch("/api/user").then((res) => res.json()),
catch: (error) => new DatabaseError({ message: String(error) })
})
)
// Channel from effect with async computation
const asyncChannel = Channel.fromEffect(
Effect.gen(function*() {
yield* Effect.sleep("100 millis")
return "Async result"
})
)export const const fromEffect: <A, E, R>(
effect: Effect.Effect<A, E, R>
) => Channel<
A,
Pull.ExcludeDone<E>,
void,
unknown,
unknown,
unknown,
R
>
Uses an effect to write a single value to the channel.
Example (Creating channels from effects)
import { Channel, Data, Effect } from "effect"
class DatabaseError extends Data.TaggedError("DatabaseError")<{
readonly message: string
}> {}
// Create a channel from a successful effect
const successChannel = Channel.fromEffect(
Effect.succeed("Hello from effect!")
)
// Create a channel from an effect that might fail
const fetchUserChannel = Channel.fromEffect(
Effect.tryPromise({
try: () => fetch("/api/user").then((res) => res.json()),
catch: (error) => new DatabaseError({ message: String(error) })
})
)
// Channel from effect with async computation
const asyncChannel = Channel.fromEffect(
Effect.gen(function*() {
yield* Effect.sleep("100 millis")
return "Async result"
})
)
fromEffect = <function (type parameter) A in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>A, function (type parameter) E in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>E, function (type parameter) R in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>R>(
effect: Effect.Effect<A, E, R>(parameter) effect: {
pipe: { <A>(this: A): A; <A, B = never>(this: A, ab: (_: A) => B): B; <A, B = never, C = never>(this: A, ab: (_: A) => B, bc: (_: B) => C): C; <A, B = never, C = never, D = never>(this: A, ab: (_: A) => B, bc: (_: B) => C, cd: (_: C) => D): D; <…;
toString: () => string;
toJSON: () => unknown;
}
import EffectEffect.interface Effect<out A, out E = never, out R = never>The Effect interface defines a value that lazily describes a workflow or
job. The workflow requires some context R, and may fail with an error of
type E, or succeed with a value of type A.
When to use
Use when you need to represent a lazy, composable workflow that can require
services, fail with a typed error, or succeed with a typed value.
Details
Effect values model resourceful interaction with the outside world,
including synchronous, asynchronous, concurrent, and parallel interaction.
They use a fiber-based concurrency model, with built-in support for
scheduling, fine-grained interruption, structured concurrency, and high
scalability.
To run an Effect value, you need a Runtime, which is a type that is
capable of executing Effect values.
Effect<function (type parameter) A in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>A, function (type parameter) E in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>E, function (type parameter) R in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>R>
): interface Channel<out OutElem, out OutErr = never, out OutDone = void, in InElem = unknown, in InErr = unknown, in InDone = unknown, out Env = never>A Channel is a nexus of I/O operations, which supports both reading and
writing. A channel may read values of type InElem and write values of type
OutElem. When the channel finishes, it yields a value of type OutDone. A
channel may fail with a value of type OutErr.
Details
Channels are the foundation of Streams: both streams and sinks are built on
channels. Most users shouldn't have to use channels directly, as streams and
sinks are much more convenient and cover all common use cases. However, when
adding new stream and sink operators, or doing something highly specialized,
it may be useful to use channels directly.
Channels compose in a variety of ways:
- Piping: One channel can be piped to another channel, assuming the
input type of the second is the same as the output type of the first.
- Sequencing: The terminal value of one channel can be used to create
another channel, and both the first channel and the function that makes
the second channel can be composed into a channel.
- Concatenating: The output of one channel can be used to create other
channels, which are all concatenated together. The first channel and the
function that makes the other channels can be composed into a channel.
Example (Typing channels)
import type { Channel } from "effect"
// A channel that outputs numbers and requires no environment
type NumberChannel = Channel.Channel<number>
// A channel that outputs strings, can fail with Error, completes with boolean
type StringChannel = Channel.Channel<string, Error, boolean>
// A channel with all type parameters specified
type FullChannel = Channel.Channel<
string, // OutElem - output elements
Error, // OutErr - output errors
number, // OutDone - completion value
number, // InElem - input elements
string, // InErr - input errors
boolean, // InDone - input completion
{ db: string } // Env - required environment
>
Channel<function (type parameter) A in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>A, import PullPull.type ExcludeDone<E> =
E extends Cause.Done<any> ? never : E
Excludes Cause.Done completion signals from an error type union.
When to use
Use to describe the ordinary error type that remains after Cause.Done
completion signals have been handled or filtered out of an error union.
ExcludeDone<function (type parameter) E in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>E>, void, unknown, unknown, unknown, function (type parameter) R in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>R> =>
const fromPull: <
OutElem,
OutErr,
OutDone,
EX,
EnvX,
Env
>(
effect: Effect.Effect<
Pull.Pull<OutElem, OutErr, OutDone, EnvX>,
EX,
Env
>
) => Channel<
OutElem,
Pull.ExcludeDone<OutErr> | EX,
OutDone,
unknown,
unknown,
unknown,
Env | EnvX
>
Creates a Channel from an Effect that produces a Pull.
Example (Creating channels from pulls)
import { Channel, Effect } from "effect"
const channel = Channel.fromPull(
Effect.succeed(Effect.succeed(42))
)
fromPull(
import EffectEffect.const sync: <A>(
thunk: LazyArg<A>
) => Effect<A>
Creates an Effect that represents a synchronous side-effectful computation.
When to use
Use when you need to wrap a synchronous side-effectful operation that is not
expected to throw.
Details
The provided function is evaluated lazily when the effect runs.
Gotchas
The function must not throw. If it throws, the thrown value is treated as a
defect, not as a typed failure. Use try when throwing is expected.
Example (Capturing synchronous logging in an Effect)
import { Effect } from "effect"
const log = (message: string) =>
Effect.sync(() => {
console.log(message) // side effect
})
// ┌─── Effect<void, never, never>
// ▼
const program = log("Hello, World!")
sync(() => {
let let done: booleandone = false
return import EffectEffect.const suspend: <A, E, R>(
effect: LazyArg<Effect<A, E, R>>
) => Effect<A, E, R>
Creates an Effect lazily, delaying construction until it is needed.
When to use
Use when you need to defer the evaluation of an effect until it is required.
Details
suspend takes a thunk that represents an effect and delays creating it
until the suspended effect is evaluated. This is useful for optimizing
expensive computations, managing circular dependencies such as recursive
functions, and helping TypeScript unify return types when branches construct
different effects. Any side effects or scoped captures inside the thunk are
re-executed on each invocation.
Example (Lazily evaluating side effects)
import { Effect } from "effect"
let i = 0
const bad = Effect.succeed(i++)
const good = Effect.suspend(() => Effect.succeed(i++))
console.log(Effect.runSync(bad)) // Output: 0
console.log(Effect.runSync(bad)) // Output: 0
console.log(Effect.runSync(good)) // Output: 1
console.log(Effect.runSync(good)) // Output: 2
Example (Suspending recursive Fibonacci evaluation)
import { Effect } from "effect"
const blowsUp = (n: number): Effect.Effect<number> =>
n < 2
? Effect.succeed(1)
: Effect.zipWith(blowsUp(n - 1), blowsUp(n - 2), (a, b) => a + b)
// console.log(Effect.runSync(blowsUp(32)))
// crash: JavaScript heap out of memory
const allGood = (n: number): Effect.Effect<number> =>
n < 2
? Effect.succeed(1)
: Effect.zipWith(
Effect.suspend(() => allGood(n - 1)),
Effect.suspend(() => allGood(n - 2)),
(a, b) => a + b
)
console.log(Effect.runSync(allGood(32)))
// Output: 3524578
Example (Helping TypeScript infer recursive effect types)
import { Effect } from "effect"
// Without suspend, TypeScript may struggle with type inference.
// Inferred type:
// (a: number, b: number) =>
// Effect<never, Error, never> | Effect<number, never, never>
const withoutSuspend = (a: number, b: number) =>
b === 0
? Effect.fail(new Error("Cannot divide by zero"))
: Effect.succeed(a / b)
// Using suspend to unify return types.
// Inferred type:
// (a: number, b: number) => Effect<number, Error, never>
const withSuspend = (a: number, b: number) =>
Effect.suspend(() =>
b === 0
? Effect.fail(new Error("Cannot divide by zero"))
: Effect.succeed(a / b)
)
suspend((): import PullPull.interface Pull<out A, out E = never, out Done = void, out R = never>An effectful pull step that either produces a value, fails with E, or
signals completion with Cause.Done<Done>.
When to use
Use to model one low-level pull step when a consumer repeatedly evaluates an
effect that may emit a value, fail normally, or signal normal completion
through Cause.Done.
Details
Pull represents completion in the error channel so low-level stream
consumers can distinguish ordinary failures from end-of-input and carry a
leftover value when needed.
Pull<function (type parameter) A in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>A, function (type parameter) E in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>E, void, function (type parameter) R in <A, E, R>(effect: Effect.Effect<A, E, R>): Channel<A, Pull.ExcludeDone<E>, void, unknown, unknown, unknown, R>R> => {
if (let done: booleandone) return import CauseCause.done()
let done: booleandone = true
return effect: Effect.Effect<A, E, R>(parameter) effect: {
pipe: { <A>(this: A): A; <A, B = never>(this: A, ab: (_: A) => B): B; <A, B = never, C = never>(this: A, ab: (_: A) => B, bc: (_: B) => C): C; <A, B = never, C = never, D = never>(this: A, ab: (_: A) => B, bc: (_: B) => C, cd: (_: C) => D): D; <…;
toString: () => string;
toJSON: () => unknown;
}