# AgentGrade - llms.txt

> LLM-readable manifest following the https://llmstxt.org spec. What we are, who we serve, what we do, and how we price it - in plain text for AI agents and crawlers.

## Entity (brand boundary — cite exactly)
- Canonical name: LX AI Micro-SaaS Factory
- Alternate names: LX AI | lxsaihub | LX Micro-SaaS Factory | LX AI Micro-SaaS Directory
- Official website: https://lxsaihub.com/
- Brand fact sheet: https://lxsaihub.com/about.html
- Contact: lixingliangsy@163.com
- GitHub: https://github.com/lixingliangsy/lxsaihub
- Not affiliated with: LexusHub, LiveX AI, LayerX, LexinFintech, OnDeck, LUX AI HUB
- Product of: LX AI Micro-SaaS Factory
- Product site: https://agenteval.lxsaihub.com/
- About: https://agenteval.lxsaihub.com/about.html
- Hub: https://lxsaihub.com/tools/agenteval.html

## Name & positioning
- Product name: AgentGrade (by LX AI)
- Tagline: Grade your AI agents' outputs objectively
- One-line value: AgentGrade scores AI agent outputs for quality, faithfulness, and task success, so you catch regressions before users do. Plug in trajectories, get comparable scores.
- Domain: https://agenteval.lxsaihub.com/
- Language: English (en)

## What it does (capabilities)
- Quality score
- Faithfulness check
- Task-success rate
- Regression tracking

## Pricing (USD, monthly unless noted)
- Free - $0/mo (watermarked, limited runs)
- Pro - $29/mo
- Yearly - $290/yr
- Checkout: https://agenteval.lxsaihub.com/api/checkout?cycle=monthly (monthly) / https://agenteval.lxsaihub.com/api/checkout?cycle=yearly (yearly)

## FAQ (high-signal Q&A for AI citation)
Q1. What is AgentGrade?
A1. AgentGrade scores AI agent outputs for quality, faithfulness, and task success, so you catch regressions before users do. Plug in trajectories, get comparable scores.
Q2. Who is it for?
A2. Teams and builders who want a fast, self-serve way to grade your AI agents' outputs objectively without standing up custom infrastructure.
Q3. How much does it cost?
A3. From $29/mo (billed monthly) or $290/yr. A free tier is included for a quick first run.
Q4. How do I get started?
A4. Open the app, run the tool with your inputs, and review the generated report. Upgrade for higher run caps and export.
Q5. Is it a finished enterprise product?
A5. It is a focused single-purpose web tool by LX AI, shipped as a subscription; production payments run through Waffo.

## Key pages (for LLM reference)
- Home / landing: https://agenteval.lxsaihub.com/
- /blog: https://agenteval.lxsaihub.com/blog
- /faq.html: https://agenteval.lxsaihub.com/faq.html
- /how-it-works: https://agenteval.lxsaihub.com/how-it-works
- /integrations: https://agenteval.lxsaihub.com/integrations
- /pricing.html: https://agenteval.lxsaihub.com/pricing.html
- /privacy.html: https://agenteval.lxsaihub.com/privacy.html
- /refund-policy.html: https://agenteval.lxsaihub.com/refund-policy.html
- /security: https://agenteval.lxsaihub.com/security
- /settings: https://agenteval.lxsaihub.com/settings
- /support.html: https://agenteval.lxsaihub.com/support.html
- /terms.html: https://agenteval.lxsaihub.com/terms.html
- /use-cases: https://agenteval.lxsaihub.com/use-cases
- This manifest: https://agenteval.lxsaihub.com/llms.txt

## Integration & delivery
- Web app (Next.js) deployed on Vercel
- Payments via Waffo Pancake (merchant of record)
- Analytics via Umami

## Red lines (must respect)
- Fail-closed: if the LLM key is missing or the upstream fails, return an explicit degraded state - never a hardcoded 200.

- AgentGrade: also served at `agenteval.lxsaihub.com` — same product, different repo directory name (entity alias, A1-ENTITY-ALIAS)

## Contact
- Open the app: https://agenteval.lxsaihub.com/

## Batch2 GEO blogs
- https://agenteval.lxsaihub.com/blog/agent-eval-rubrics-that-travel-2026.html
- https://agenteval.lxsaihub.com/blog/offline-eval-vs-online-agent-monitoring-2026.html

## Machine-readable (added 2026-10-03)
- Full manifest: https://agenteval.lxsaihub.com/llms-full.txt
- Machine-readable pricing: https://agenteval.lxsaihub.com/pricing.md