# pattern.yaml
name: content-moderation
version: 1.0.0
agents:
capabilities: [moderation]
min: 3
execution:
strategy: parallel
aggregation:
strategy: voting
method: majority
validation:
minConfidence: 0.8
output:
verdict: $vote.result
confidence: $vote.confidence
# Run and get a confidence-scored result
parallax run pattern.yaml --input '{ "content": "..." }'
Why Parallax?
Orchestrate reliable multi-agent systems with measurable confidence.
Orchestration Patterns
Compose consensus, voting, merge, and quality-gate flows with simple YAML or the visual builder.
Consensus & Voting
Aggregate multiple agents into a single result with traceable decisions and confidence scores.
Confidence Scoring
Quantify uncertainty and enforce thresholds before results move downstream.
Agent Runtime Control
Spawn, monitor, and shut down CLI agents locally, in Docker, or in Kubernetes.
Observability
Trace executions end-to-end with logs, metrics, and audit-ready event history.
Production-Ready
Scale to HA clusters with persistence, scheduling, and enterprise security.
Built for real workloads
Turn brittle agent calls into resilient systems.
Content Moderation
Cross-check multiple models, vote, and block low-confidence results.
Code Review
Parallelize specialized agents and merge findings into a single report.
Data Extraction
Combine multiple interpretations and validate against a confidence threshold.
QA & Verification
Route tasks through quality gates before they reach production.
Start orchestrating in minutes
Install the SDK, run your first pattern, and scale as you grow.