Production Patterns
End-to-end examples showing how ElectriPy components compose into production-ready AI systems. Each recipe is a runnable, commented Python file.
Observable Agent Routing
Compose workload routing with observability and policy evaluation. Route requests to the right provider, gate actions with policy checks, and trace the full flow.
When to use
When you need provider routing with fallback, runtime policy checks, and end-to-end tracing.
LLM Gateway
Provider-agnostic LLM gateway with sync/async support. Wraps OpenAI, Anthropic, Ollama, and HTTP-JSON providers behind a unified interface.
When to use
When you want to swap LLM providers without changing application code.
Policy-Governed Execution
Runtime policy evaluation with typed decisions, violation reporting, approval workflows, and evidence requirements. Gate any agent action with composable rules.
When to use
When you need governance, compliance enforcement, or approval workflows for agent actions.
Evaluation Pipeline
Evaluate retrieval quality, answer correctness, and tool-call accuracy with pluggable scorers. Generate CI-friendly reports for regression gates.
When to use
When you need systematic evaluation of AI pipelines before or after deployment.
AI Telemetry & Observability
Instrument AI workloads with structured spans, redaction, and OpenTelemetry export. Capture latency, token usage, model metadata, and span-level attributes.
When to use
When you need visibility into cost, latency, and quality across agent workloads in production.
Agent Collaboration
Multi-agent coordination patterns. Orchestrate specialist agents with shared context, handoff protocol, and result aggregation.
When to use
When a single LLM call is insufficient and you need multi-agent decomposition.
CLI Tool
Build typed CLI tools with ElectriPy's CLI framework. Subcommands, typed arguments, help generation, and config loading wired together.
When to use
When you need a production-grade CLI over your AI pipeline or toolkit.