prowl.world
TOPIC concept hub

ASO: Agent Search Optimization

SEO was about being visible to humans through search engines. ASO is about being usable by autonomous agents through tool registries, structured docs, and benchmark networks. The discipline is two years old, has eight measurable dimensions, and is currently the cheapest way for an API vendor to capture share of the agent economy.

The one-line definition

Agent Search Optimization (ASO) is the practice of making an API or service discoverable, parseable, and benchmarkable by autonomous AI agents.

The 8 dimensions

Across hundreds of API benchmarks, eight dimensions explain most of the variance in agent success. These are what Prowl measures, and what you should optimize.

DimensionWeightWhat it measures
Token efficiency25%Tokens needed to understand and use the API
First-try success20%% of calls that succeed without retry
Response parseability15%JSON consistency, no HTML errors, stable shapes
Error clarity15%Machine-readable codes, actionable messages
Doc quality10%OpenAPI completeness, llms.txt, examples
Auth simplicity5%Steps from credential to first call
Latency5%Raw response speed
Consistency5%Same request → same response shape

The 6 highest-leverage tactics

  1. Publish a /llms.txt at your API root. 1,000-3,000 token markdown summary with endpoints, auth, examples. Agents fetch this first.
  2. Submit to MCP registries — Official Registry, Smithery, PulseMCP, MCP.directory. Each is under an hour.
  3. Audit error responses — JSON only, machine-readable codes, actionable messages.
  4. Get benchmarkedpip install prowl-bench && prowl-bench run https://your-api.com. Free, takes 5 minutes, gives you a prioritized fix list.
  5. Add JSON-LD to your homepage — Organization, WebSite, SoftwareApplication schemas. LLM search engines preferentially quote structured data.
  6. Write a definitional post for your category. Becomes the LLM-citation asset for years.

Read more

Get your ASO score in 5 minutes.
Run a free benchmark → pip install prowl-bench