Actor-based in-memory cache with TTL and LRU eviction for Swift 6.
SwiftCache is a lightweight, thread-safe key-value cache built with Swift 6 strict concurrency. No external dependencies — just Foundation.
- Actor-isolated — all reads and writes are
async, safe to call from any concurrency domain - TTL support — per-entry or global time-to-live; expired entries are evicted lazily
- LRU eviction — least-recently-used entries are dropped when capacity is exceeded
getOrInsert— compute-and-cache in a single call, no double-checked locking needed- Statistics — track hit/miss counts and hit rate
CacheKeyhelpers — namespaced and composite key builders
// Package.swift
dependencies: [
.package(url: "https://github.com/VDurocher/SwiftCache.git", from: "1.0.0")
]import SwiftCache
// Create a cache with a 5-minute TTL and max 256 entries
let cache = Cache<String, Data>(capacity: 256, defaultTTL: 300)
// Store
await cache.set("avatar-42", value: imageData)
// Retrieve
let data = await cache.get("avatar-42") // Data? — nil if missing or expired
// Compute-and-cache
let profile = try await cache.getOrInsert("profile-7") {
try await api.fetchProfile(id: 7)
}
// Namespaced keys
let key = CacheKey.namespaced("user", id: 42) // "user:42"// Read
func get(_ key: Key) -> Value?
// Write
func set(_ key: Key, value: Value, ttl: TimeInterval? = nil)
// Compute-and-cache
func getOrInsert(_ key: Key, ttl: TimeInterval? = nil,
compute: @Sendable () async throws -> Value) async rethrows -> Value
// Removal
func remove(_ key: Key)
func purgeExpired()
func flush()
// Metadata
var count: Int
var statistics: CacheStatistics // hits, misses, hitRateMIT — @VDurocher
