Prowl
62/100
prowl
Benchmarked Sep 25, 2026

Smartphone Test Farm

Control and manages real Smartphone devices from browser and restful apis

api platform_profile
Benchmark Your API

Score Breakdown

Parseability8/10
Token Efficiency7/10
Consistency6/10
Documentation6/10
Error Clarity6/10
Auth Simplicity6/10
Latency5/10
First-Try Success5/10

Benchmark Analysis Log

Full LLM thinking from the 4-phase benchmark pipeline.

Analyze
{
  "service_type": "platform",
  "base_url": "http://openstf.io",
  "auth_method": "none",
  "auth_config": {},
  "endpoints": [],
  "pricing_model": {
    "type": "free",
    "details": {
      "note": "OpenSTF is an open-source project, free to use and self-host."
    }
  },
  "rate_limits": {},
  "capabilities": [
    "Remote control of real Android devices via browser",
    "RESTful API for device management and automation",
    "Device reservation and sharing",
    "Screen mirroring and interaction",
    "ADB over network",
    "Device monitoring and health checks",
    "Support for multiple users and teams"
  ],
  "raw_analysis": "OpenSTF (Smartphone Test Farm) is an open-source platform that allows users to control and manage real Android devices remotely from a web browser. It provides a web interface and RESTful APIs for device interaction, making it suitable for mobile testing, development, and debugging. The platform is designed for developers, QA engineers, and teams needing access to physical devices. It is self-hosted and free to use. The website (openstf.io) appears to be a WordPress-based blog or landing page, with no direct API documentation available. The actual API is part of the self-hosted application, typically documented in the project's GitHub repository (https://github.com/openstf/stf). The platform is mature, with a large community and commercial support available through other vendors. Integrations include ADB, various CI systems, and custom scripts via REST API."
}
Execute

3/3 tests passed

TestEndpointStatusLatency
website_uptimeGET /2001004ms
robots_txtGET /robots.txt2001323ms
llms_txtGET /llms.txt2001431ms
Interpret
```json
{
  "overall": 58,
  "dimensions": {
    "token_efficiency": 7.0,
    "first_try_success": 5.0,
    "response_parseability": 8.0,
    "error_clarity": 5.5,
    "doc_quality": 6.0,
    "auth_simplicity": 6.0,
    "latency": 5.0,
    "consistency": 5.5
  },
  "pricing_normalized": {
    "model": "free",
    "notes": "Open-source, free to self-host. No managed SaaS — user bears hosting, device provisioning, and ops cost."
  },
  "issues": [
    "robots.txt and llms.txt both return HTML (the homepage), not actual robots/llms directives — misconfigured routing, so agents can't discover crawl rules or agent-friendly docs.",
    "No security headers detected on any response; weak signal for production readiness.",
    "Latency ~1.0–1.4s per request — usable but not snappy for agentic loops.",
    "Self-hosted-only model means 'onboarding' actually means standing up infrastructure (Docker, ADB, device farm), which is far from minutes-to-productive.",
    "Website appears to be a WordPress marketing page; the actual product docs/API live elsewhere (GitHub), fragmenting the description agents must give users.",
    "No managed cloud offering or hosted API endpoint — agents can't just point users at a key and go."
  ],
  "recommendations": [
    "Fix robots.txt and llms.txt to return real plain-text directives (or at least valid content) so agent tooling can parse crawl/agent guidance instead of hitting the HTML homepage.",
    "Add a concise /llms.txt summarizing capabilities, self-host steps, REST API base, and ADB-over-network usage in machine-readable form.",
    "Publish a quickstart with a docker-compose one-liner and default credentials, plus an API reference (device list, reserve, connect) with example curl calls — this would sharply improve first_try_success.",
    "Add security headers (HSTS, CSP, X-Content-Type-Options) to signal production quality and improve trust scoring.",
    "Clarify limitations in docs: device compatibility matrix, ADB requirement, port/firewall needs, and concurrency limits.",
    "Consider a hosted/managed tier or a reference cloud deployment to remove the self-host tax and raise auth_simplicity from manual user provisioning to a simple API key or SSO."
  ]
}
```

Agent Readiness

x402 Payments
Not supported
Streaming
No
Sandbox
None
Agent Auth
Unknown
SDKs
None listed
MCP Support
No

Embed your Prowl badge

Show your live agent-readiness score on your own site. Free, no auth — it updates as your score changes.

Prowl agent-readiness badge
<a href="https://prowl.world/service/smartphone-test-farm">
  <img src="https://prowl.world/badge/smartphone-test-farm.svg" height="56" alt="Agent-readiness on Prowl">
</a>

Options: ?style=light|dark · ?size=sm|md · ?variant=certified (claimed + DNS-verified only) · badge generator with preview

Want the full interactive view?

See operational metrics, LLM evaluations, agent readiness, and more.

Open in Dashboard