Agent-ready public evidence

ToolVitals is built for agents, too.

Every public tool profile has a clean Markdown representation and structured JSON endpoint. Agents can search, shortlist by use case, compare tools, inspect trends, and retain one release snapshot across the workflow.

Current machine snapshot

One identity across every surface

Responses expose the same snapshot identity so clients can detect mixed or stale evidence before making a consequential comparison.

Snapshot
tv-20260911-8ab5e3f1311b
Public profiles
3,392
Discovery profiles
2,760
Generated
2026-09-11T07:01:56.000Z

Discovery

Stable entrypoints

Read-only retrieval

Start small, then inspect evidence

Filtered endpoints avoid loading multi-megabyte bulk indexes into an agent context. Limits are bounded and general discovery excludes noindex profiles.

Search/api/agent/search?q=postgres&limit=10
Tool JSON/api/agent/tools/net
Tool Markdown/tools/net.md
Use-case shortlist/api/agent/use-cases/database-systems?limit=10
Compare/api/agent/compare?a=postgresql&b=mariadb
Trends/api/agent/trends/postgresql?days=90

Trust boundary

What agents must not infer

  • Scores summarize bounded public maintenance, shipping, openness, and evidence signals. They are not user-review ratings.
  • Null means not collected or not confidently established—not zero.
  • Popularity is adoption context, not evidence of quality.
  • Public signals do not prove security, uptime, support quality, customer satisfaction, or production suitability.
  • Sponsorship, affiliate relationships, and commercial activity do not affect ToolVitals scores, rankings, or endpoint ordering.

Tool profiles and research are generated from automated public-signal collection and source synthesis. The profile itself shows confidence, freshness, and source receipts so people and agents can verify material claims.