MCP Server

Straden exposes a Model Context Protocol server, so AI assistants and IDEs such as Claude Code, Cursor or VS Code can write k6 scripts, run them and analyse the results against your Straden instance.

Endpoint

POST https://<your-straden-domain>/mcp
Authorization: Bearer <token>

The server uses streamable HTTP, is authenticated with Laravel Sanctum and is rate limited to 120 requests per minute per user.

Create a token

Go to Settings → API tokens, name the token, choose an expiry (30, 90, 365 days or never) and select its abilities:

Ability	Grants
`mcp:read`	Read projects, scripts, runs and metrics
`mcp:write`	Create and edit scripts
`mcp:run`	Start and cancel runs

The token is shown once — copy it straight into your client config.

Client configuration

Claude Code

BASH
claude mcp add --transport http straden https://straden.example.com/mcp \
  --header "Authorization: Bearer $STRADEN_TOKEN"

Generic JSON config (Cursor, VS Code, …)

JSON
{
  "mcpServers": {
    "straden": {
      "url": "https://straden.example.com/mcp",
      "headers": { "Authorization": "Bearer <token>" }
    }
  }
}

Tools

Projects & scripts

Tool	Ability	Description
`list-projects`	read	Projects and their tests
`get-test-context`	read	Target URL, connectors, scripts and recent runs for a test
`create-script`	write	Create a new script for a test
`update-script`	write	Replace a script's entry point
`validate-script`	read	`k6 inspect` + structural checks
`list-script-files` / `read-script-file`	read	Browse a script folder
`write-script-file` / `rename-script-file` / `delete-script-file`	write	Edit a script folder
`get-script-insights`	read	Recent run results for a script

Runs

Tool	Ability	Description
`start-run`	run	Queue a run for a script
`get-run-status`	read	Poll a run's status
`cancel-run`	run	Cancel a queued or running run
`list-runs`	read	Runs for a test or script
`get-run-context`	read	Summary metrics, thresholds and config
`get-run-metrics`	read	InfluxDB time-series for a run
`read-run-log`	read	Raw k6 console output
`get-run-insight`	read	The AI insight report, if generated

Observability

Tool	Description
`prometheus-query` / `prometheus-query-range`	PromQL against an attached Prometheus connector
`prometheus-list-metrics` / `prometheus-metadata`	Discover available metrics
`database-metrics`	PostgreSQL / MySQL / MongoDB statistics
`redis-metrics`	Redis server statistics

Typical workflow

list-projects, then get-test-context for the target URL, connectors and existing scripts.
create-script (or update-script / write-script-file), then validate-script and fix any issues.
start-run, then poll get-run-status every ~10 seconds until it's passed, failed, aborted or error.
Analyse with get-run-context, get-run-metrics, read-run-log and get-script-insights; correlate with the Prometheus, database and Redis tools.
Fix the script or the application under test, and repeat.