Clarvia: AEO quality scores for API/function call readiness

[digitamaz]

Clarvia — API/Agent Readiness Scoring for Function Calls

Hi Gorilla team 👋

Gorilla's mission is training and evaluating LLMs for function calls. We've built Clarvia (https://clarvia.art), a complementary tool that scores APIs and MCP servers for agent-engine optimization (AEO) — essentially measuring how well a service is designed to be called by AI agents.

What Clarvia provides

• AEO Score (0–100): Measures documentation quality, schema completeness, error handling, auth clarity, and rate limit transparency — exactly what matters for LLM function calling
• 27,000+ indexed tools: Every major API and MCP server already scored
• MCP server (npm: ): Agents can query scores and find alternatives before making calls
• Gate-check tool: Validate tool readiness before adding to agent workflows

Why this is relevant to Gorilla

Gorilla evaluates LLM accuracy on function calls — but the quality of the API documentation directly affects LLM performance. Clarvia's AEO scores could be a useful signal for:

• Selecting which APIs to include in Gorilla's training/eval sets (higher AEO = cleaner docs = better LLM comprehension)
• Filtering out poorly-documented APIs that introduce noise
• Benchmarking API documentation quality alongside function call accuracy

Integration idea

Using Clarvia MCP tools to pre-screen APIs before adding them to eval sets:

npx clarvia-mcp-server
# Tools: search_tools, aeo_score, gate_check, get_alternatives

API: https://clarvia-api.onrender.com/docs

Would love to discuss how Clarvia's quality scores could complement Gorilla's function call evaluations.

— Clarvia team