plano/demos
Tim e448479d3b demos: add peyeeye PII redaction & rehydration filter chain
Mirrors the pii_anonymizer demo: a FastAPI input/output filter pair that
delegates PII detection to the Peyeeye API (https://api.peyeeye.ai).

- /redact/{path:path}    redacts user text in chat completions, Anthropic
                          messages, and Responses-API request bodies.
- /rehydrate/{path:path} restores placeholders in the LLM response, then
                          best-effort deletes the stateful session.

Fail-closed semantics: any unexpected response shape from /v1/redact, or
a length mismatch between sent and returned text counts, raises a 502
rather than forwarding partially-redacted data. Auth errors map to 500.

Stateful (default) and stateless (sealed skey_... blob) session modes
are both supported via PEYEEYE_SESSION_MODE.

Includes 24 pytest cases covering the round trip, length-guard,
unexpected-shape guard, multimodal content lists, Anthropic and
Responses formats, and stateless mode.
2026-04-25 23:00:23 -05:00
..
advanced fix: resolve all open Dependabot security alerts (#866) 2026-04-02 18:00:28 -07:00
agent_orchestration fix: resolve all open Dependabot security alerts (#866) 2026-04-02 18:00:28 -07:00
filter_chains demos: add peyeeye PII redaction & rehydration filter chain 2026-04-25 23:00:23 -05:00
getting_started Update black hook for Python 3.14 (#857) 2026-03-31 13:18:45 -07:00
integrations Run demos without Docker (#809) 2026-03-11 12:49:36 -07:00
llm_routing Bump version to 0.4.22 (#917) 2026-04-24 16:43:19 -07:00
shared ci(e2e): stabilize preference demo test execution (#865) 2026-04-02 21:32:20 -04:00
README.md Support for Codex via Plano (#808) 2026-03-10 20:54:14 -07:00

Plano Demos

This directory contains demos showcasing Plano's capabilities as an AI-native proxy for agentic applications.

Getting Started

Demo Description
Weather Forecast Core function calling with a weather query agent, interactive chat UI, and Jaeger tracing
LLM Gateway Key management and dynamic routing to multiple LLM providers with header-based model override

LLM Routing

Demo Description
Preference-Based Routing Routes prompts to LLMs based on user-defined preferences and task type (e.g. code generation vs. understanding)
Model Alias Routing Maps semantic aliases (arch.summarize.v1) to provider-specific models for centralized governance
Claude Code Router Extends Claude Code with multi-provider access and preference-aligned routing for coding tasks
Codex Router Extends Codex CLI with multi-provider access and preference-aligned routing for coding tasks

Agent Orchestration

Demo Description
Travel Agents Multi-agent travel booking with weather and flight agents, intelligent routing, and OpenTelemetry tracing
Multi-Agent CrewAI & LangChain Framework-agnostic orchestration combining CrewAI and LangChain agents in unified conversations

Filter Chains

Demo Description
HTTP Filter RAG agent with filter chains for input validation, query rewriting, and context building
MCP Filter RAG agent using MCP-based filters for domain validation, query optimization, and knowledge base retrieval

Integrations

Demo Description
Ollama Use Ollama as a local LLM provider through Plano
Spotify Bearer Auth Bearer token authentication for third-party APIs (Spotify new releases and top tracks)

Advanced

Demo Description
Currency Exchange Function calling with public REST APIs (Frankfurter currency exchange)
Stock Quote Protected REST API integration with access key management
Multi-Turn RAG Multi-turn conversational RAG agent for answering questions about energy sources
Model Choice Test Harness Evaluation framework for safely testing and switching between models with benchmark fixtures