Zenstack

{
  "service_type": "platform",
  "base_url": "https://zenstack.dev",
  "auth_method": "none",
  "auth_config": {},
  "endpoints": [],
  "pricing_model": {
    "type": "freemium",
    "details": {
      "free_tier": "Open-source core available under MIT license",
      "commercial": "Optional paid plans for enterprise features and support"
    }
  },
  "rate_limits": {},
  "capabilities": [
    "Type-safe ORM for TypeScript",
    "Built-in access control (RBAC/ABAC via declarative policies)",
    "Automatic query services (REST, GraphQL, tRPC)",
    "Prisma schema integration",
    "Data validation and modeling",
    "Policy engine for authorization",
    "Code generation for CRUD operations",
    "Serverless and edge-ready",
    "Database support: PostgreSQL, MySQL, SQLite, SQL Server, MongoDB",
    "Framework integrations: Next.js, Express, Remix, Nuxt, Better Auth, etc."
  ],
  "raw_analysis": "ZenStack is not a conventional API-first SaaS tool but a developer framework/platform — a modern data layer for TypeScript applications. It combines a type-safe ORM (built on Prisma) with declarative access control and automatic CRUD query service generation. This is a library/package, not a hosted service, so there are no public REST API endpoints to interact with. Instead, developers integrate ZenStack into their own codebases by installing npm packages (@zenstackhq/*) and defining schemas. Authentication is handled by the host application (e.g., via NextAuth, Clerk, etc.), and ZenStack's policy engine enforces authorization based on the user context passed to it. As an open-source project (core under MIT), it has a freemium model with optional commercial subscriptions for enterprise-grade features and support. It is targeted at TypeScript developers building full-stack applications who need fine-grained access control and automatic API generation without writing boilerplate. The platform is mature and actively maintained, with a growing ecosystem of plugins and integrations (e.g., for Next.js, Express, and various databases). While it lacks a public REST API, it exposes its functionality through generated query services and a plugin system, making it a powerful internal tool rather than an external API service."
}

2/3 tests passed

{
  "overall": 76,
  "dimensions": {
    "token_efficiency": 8.0,
    "first_try_success": 8.5,
    "response_parseability": 8.0,
    "error_clarity": 7.5,
    "doc_quality": 8.5,
    "auth_simplicity": 9.0,
    "latency": 9.5,
    "consistency": 8.0
  },
  "pricing_normalized": {
    "model": "freemium",
    "free_tier": "Open-source core under MIT; paid enterprise tier for extra features/support"
  },
  "issues": [
    "No /llms.txt endpoint returns 404, which hurts machine-readable discoverability for agents",
    "Auth is only via code/SDK integration — there is no hosted identity flow (no notlogin-style agent onboarding)",
    "Docs site is on Docusaurus v3.4.0 and marketing on Framer, so content is split across two stacks",
    "Error docs around policy evaluation edge cases are likely to be thin for an authorization-focused product"
  ],
  "recommendations": [
    "Publish an /llms.txt index of docs and API surfaces to improve agent ingestion",
    "Provide an OpenAPI or schema endpoint that agents can fetch to programmatically explore the query-service API",
    "Add explicit code samples for each framework integration to boost first-try success",
    "Document error codes and policy denial reasons in a structured table",
    "Consider a `create-zenstack-app` scaffold with a one-command demo to compress time-to-value",
    "Expose a status page/uptime endpoint for agents to verify reliability"
  ]
}