Feature flags engineered for clean code and zero PII leakage.
Sub-microsecond in-process evaluations, built-in stale flag hygiene, and peer-reviewed change governance — without the per-MAU billing trap.
package main
import (
"context"
"fmt"
flagura "github.com/dhawalhost/flagura/sdks/go"
)
func main() {
// In-memory local evaluation (~85ns latency)
client := flagura.NewClient("https://api.flagura.dev", "flg_live_secret",
flagura.WithLocalEvaluation(true),
)
defer client.Close()
res, _ := client.Evaluate(context.Background(), "ai-smart-search", flagura.Context{
UserID: "usr_alex_42",
Email: "alex@company.com",
})
fmt.Printf("Enabled: %v | Latency: %v ns\n", res.Enabled, res.EvaluationLatencyNs)
}import { FlaguraClient } from '@flagura/sdk';
// Instant local in-memory evaluation over persistent SSE stream
const client = new FlaguraClient({
endpoint: 'https://api.flagura.dev',
apiKey: 'flg_live_secret'
});
const { enabled, variant } = await client.evaluate('ai-smart-search', {
user_id: 'usr_alex_42',
email: 'alex@company.com',
});
// Evaluated result: { enabled, variant }from flagura import FlaguraClient, EvaluationContext
# Fast local in-process evaluation with zero remote HTTP calls
client = FlaguraClient("https://api.flagura.dev", api_key="flg_live_secret")
res = client.evaluate("ai-smart-search", EvaluationContext(
user_id="usr_alex_42",
email="alex@company.com"
))
print(res.enabled, res.variant)use flagura::{FlaguraClient, EvaluationContext};
#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
let client = FlaguraClient::new("https://api.flagura.dev")
.with_api_key("flg_live_secret");
let ctx = EvaluationContext::new("usr_alex_42").with_email("alex@company.com");
if client.is_enabled("ai-smart-search", &ctx).await {
println!("✨ AI Smart Search is active!");
}
Ok(())
}curl -X POST https://api.flagura.dev/api/v1/evaluate \
-H "Content-Type: application/json" \
-H "Authorization: Bearer flg_live_secret" \
-d '{"flags":["ai-smart-search"],"context":{"user_id":"usr_alex_42","email":"alex@company.com"}}'Built for clean codebases and operational peace of mind.
Why engineering organizations choose Flagura over legacy SaaS and complex multi-container stacks.
Zero Flag Debt & Dead Flag Graveyard Cleanup
Automated detection of 100% rolled-out flags
The silent killer of feature flagging is code rot: dead flags lingering in codebases for years because nobody knows if deleting them is safe. Flagura continuously audits flags, detects features at 100% rollout, tracks longevity, and guides teams to clean up dead code.
4-Eyes Governance
PR-Style Release Approvals
Require senior peer review and two-person sign-off before dangerous flags or kill switches can be toggled in production environments.
Zero PII Leakage
Air-Gapped & Self-Hosted
Flagura sends compiled targeting schemas to your application. Customer emails, IP addresses, and user IDs never leave host process memory—no user PII leaves your infrastructure.
~85ns Local Engine
Zero Network Roundtrips
Evaluated in CPU L1/L2 cache with 64-bit FNV-1a deterministic sticky hashing. 0 heap allocations, 0 blocking HTTP calls, and zero single points of network failure on critical user paths.
Auto-Tripwire Breaker
P99 Error Budget Protection
Integrated health monitoring triggers automatic canary rollback within milliseconds if downstream HTTP error rates or crash exceptions spike during a release.
Predictable Infrastructure • No MAU Billing Penalties
Never get punished for user growth. Evaluate 10,000 requests or 10 billion requests at flat, predictable cost.
How updates travel to every edge node.
Flags and rollout strategies are persisted in Embedded SQLite or PostgreSQL and signed cryptographically.
Changes are published across persistent Server-Sent Events (SSE) connections with auto-reconnect backoff.
Local SDK memory caches update atomically with RWMutex zero-allocation fast paths.
In-Browser Real-Time Benchmark Runner
Execute 100,000 deterministic flag evaluations directly in your browser JavaScript runtime.
curl -sSL https://flagura.dev/install.sh | bashgo install github.com/dhawalhost/flagura/cmd/cli@latestdocker run --rm -it ghcr.io/dhawalhost/flagura:latest flagura --help