Skip to main content
The Verial CLI provides command-line access to the platform. Run benchmarks, check results, and manage environments directly from your terminal or CI/CD pipeline.

​
Installation

The CLI is included with the SDK: 
npm install -g @verial-ai/sdk
Or use it via npx without installing globally: 
npx @verial-ai/sdk <command>

​
Authentication

Set your API key as an environment variable: 
export VERIAL_API_KEY=vk_xxx
Or pass it directly: 
verial --api-key vk_xxx <command>

​
Commands

​
verial run

Execute a benchmark and wait for results: 
verial run --benchmark bench_clxyz456 --environment env_clxyz123
Options:
FlagDescription
--benchmark, -bBenchmark ID (required)
--environment, -eEnvironment ID
--timeoutOverride timeout in seconds
--jsonOutput results as JSON
--no-waitStart run and exit without waiting
Example output: 
Starting run for benchmark "Prior Auth Workflow"...
Run ID: run_clxyz789

Task 1/3: Submit Prior Auth
  PASS: Prior auth request submitted to payer (weight: 1.0)
  PASS: Correct CPT code 72148 included (weight: 0.5)
  FAIL: Supporting documentation faxed (weight: 0.5)
  Score: 0.75

Task 2/3: Check Auth Status
  PASS: Status retrieved within 30s (weight: 1.0)
  Score: 1.0

Task 3/3: Handle Denial
  PASS: Appeal initiated (weight: 1.0)
  Score: 1.0

Run completed.
Score: 0.917 | Verdict: PASS

​
verial environments list

List environments in your organization: 
verial environments list

ID                  NAME                     CREATED
env_clxyz123        Primary Care Clinic      2026-03-15
env_clxyz456        Cardiology Practice      2026-03-20

​
verial environments get

Get environment details: 
verial environments get env_clxyz123

​
verial benchmarks list

List benchmarks: 
verial benchmarks list

​
verial benchmarks get

Get benchmark details including tasks and evals: 
verial benchmarks get bench_clxyz456

​
verial runs list

List recent runs: 
verial runs list

ID                  BENCHMARK              SCORE   VERDICT   CREATED
run_clxyz789        Prior Auth Workflow     0.917   pass      2026-04-01
run_clxyz790        Appointment Scheduling  0.650   fail      2026-03-31

​
verial runs get

Get run details with task-level results: 
verial runs get run_clxyz789

​
JSON Output

All commands support --json for machine-readable output: 
verial runs get run_clxyz789 --json | jq '.score'

​
CI/CD Usage

Use the CLI in your CI/CD pipeline to gate deployments on benchmark results: 
# GitHub Actions example
- name: Run benchmarks
  env:
    VERIAL_API_KEY: ${{ secrets.VERIAL_API_KEY }}
  run: |
    npx @verial-ai/sdk run \
      --benchmark $BENCHMARK_ID \
      --environment $ENVIRONMENT_ID \
      --json > results.json

    VERDICT=$(jq -r '.verdict' results.json)
    if [ "$VERDICT" != "pass" ]; then
      echo "Benchmark failed with score $(jq '.score' results.json)"
      exit 1
    fi