Unkey
One place to deploy an API, issue and revoke API keys for it, rate-limit it, and watch its logs — instead of stitching together four separate services to do each job.
🔗 Visit Unkey
Description
Shipping a production API usually means assembling a stack from several different tools: something to deploy the code, something else to issue and manage API keys, another service for rate limiting, and a fourth for logs and analytics. Unkey bundles all of that into one platform, so a developer can go from a Git push to a deployed, authenticated, rate-limited, and observable API without gluing together separate vendors.
Unkey is open-source (AGPL) API infrastructure written in Go, offering Git-based automatic deployment with Docker support, global edge gateways for rate limiting and access control, API key issuance and instant revocation, and built-in observability with automatic log collection. It uses a pay-per-execution compute model — you're charged for vCPU time only while your code actually runs, not during idle periods — plus preview environments with a 30-minute rollback window and OpenAPI-based validation to catch breaking changes before they ship.
💬 Our review
The short version: if you're currently duct-taping together a deployment platform, an API key service, and a rate limiter, Unkey's all-in-one approach genuinely simplifies the stack — and being open source with a free 150K-requests tier means you can try it without committing to a vendor.
Against point solutions — Vercel or Render for deployment, a separate service like Kong for API management — Unkey's bet is that most teams don't need best-in-class everything, they need one thing that's good enough across deployment, keys, and observability, with less integration overhead. The tradeoff is that a bundled platform is inherently a jack-of-all-trades: teams with very specific gateway or deployment requirements might still get more control from specialized tools. Pricing scales reasonably with usage (free to $1000/month based on volume), which keeps it approachable for early-stage APIs while giving room to grow, though it's worth checking the compute pricing model carefully for very bursty workloads.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Gestion API : gratuit jusqu'à 150K requêtes/mois, Pro de 25 $ à 1000 $/mois selon le volume (250K à 100M requêtes/mois). Deploy : 5 $ à 50 $/mois de base + usage compute (0,000006944 $/vCPU-seconde) + frais de sortie réseau.
Pros
Open source (AGPL), auditable et transparent
Regroupe déploiement, clés API, rate limiting et observabilité en une seule plateforme
Modèle de facturation à l'exécution — pas de coût pendant les périodes d'inactivité
Portail développeur en libre-service pour la gestion des clés
Cons
Solution généraliste — moins de contrôle fin que des outils spécialisés type Kong
Modèle de coût compute à vérifier attentivement pour des charges très irrégulières
Palier Pro peut monter jusqu'à 1000 $/mois pour un fort volume
Date de fondation non communiquée