How do you want to test?

Run it yourself in the browser, or hand the free checks to your coding agent. Both read the same checklists and the same rules.

Test it yourself

The tester, in your browser. Pick a site, pick a checklist, and see where the agent behaves and where it does not.

  • Free pre-flight checks before anything is spent
  • Demo mode with three fake models, no key needed
  • Real models with your own key and router; the cost is shown first
  • A report with the evidence for every verdict
Open the tester

Connect an agent

The zero-token checks, for coding agents: a command-line tool and a local MCP server. No key, nothing spent.

  • Lint a bundle and size each agent's call
  • Validate a single tool call before it is made
  • List checklists and estimate a run's cost
  • Judge a saved trace against a checklist
Show me how

For AI agents

Connect your agent

The zero-token checks need no key and spend nothing. Run them from a clone of this repository.

On the command line

node cli/tools.mjs lint-bundle path/to/bundle
node cli/tools.mjs call-size path/to/bundle
node cli/tools.mjs validate-call path/to/bundle --tool website_add_to_cart --args '{"id":"margherita","quantity":1}'
node cli/tools.mjs list-checklists --industry restaurants
node cli/tools.mjs estimate path/to/bundle --checklist restaurants-menu.v1
node cli/tools.mjs check-trace trace.json --checklist restaurants-menu.v1 --dir path/to/bundle

Exit codes: 0 passed, 1 a check found a problem, 2 bad input. Output is JSON.

As a local MCP server

The same checks, for coding agents. Keys never leave your machine.

{
  "mcpServers": {
    "agentic-stress-test": {
      "command": "node",
      "args": ["path/to/agentic-stress-test/mcp/server.mjs"]
    }
  }
}

Runs that spend model calls are not offered to agents. They stay on the page and on the command line, behind the cost check. The site's own description for agents is at llms.txt.