{
  "@context": "https://schema.org",
  "@type": "SoftwareApplication",
  "@id": "https://www.insightits.com/products/chorusgraph.html",
  "slug": "chorusgraph",
  "canonicalName": "ChorusGraph",
  "name": "ChorusGraph — Native Agent Runtime",
  "canonicalUrl": "https://www.insightits.com/products/chorusgraph.html",
  "knowledgeUrl": "https://www.insightits.com/catalog/chorusgraph.json",
  "markdownUrl": "https://www.insightits.com/catalog/chorusgraph.md",
  "family": "agentRuntime",
  "infraFamily": "Agent Runtime",
  "recordType": "product",
  "status": "published",
  "parentProduct": null,
  "whatItIs": "Apache-2.0 native Python agent runtime and LangGraph alternative with its own Graph engine, semantic cache, Route Ledger, memory, and ports for PrismGuard security and PrismRAG retrieval. It does not require LangGraph in production.",
  "problem": "Production agents fail on glue: repeat-intent LLM spend without a cache, logs instead of decision replay, and six repos duck-taped (orchestration + Redis + retrieval + checkpoints + audit) before go-live.",
  "isNot": [
    "A LangGraph wrapper or plugin",
    "A claim that LangGraph has no cache — LangGraph has opt-in CachePolicy / compile(cache=...)",
    "A published performance comparison against CrewAI or Microsoft Agent Framework"
  ],
  "alternatives": [
    {
      "name": "LangGraph",
      "relationship": "Published Azure scale comparison uses LangGraph as a documented baseline. ChorusGraph is a native runtime with cache, Route Ledger, and ports included; LangGraph is a mature ecosystem teams compose themselves.",
      "url": "https://www.insightits.com/compare/chorusgraph-vs-langgraph.html"
    }
  ],
  "features": [
    {
      "name": "+4–15 pp Task Success",
      "description": "Published Azure scale run heavy_20260708_140300 (n=300 per scenario) vs the documented LangGraph baseline."
    },
    {
      "name": "300+ Deterministic Tests",
      "description": "CI runs without live API keys; benchmark methodology and raw artifacts are published separately."
    },
    {
      "name": "Security + Retrieval Plug-ins",
      "description": "PrismGuard (make_guard_handler) and PrismRAG (your DB taxonomy) — legal, healthcare, finance without forking the engine."
    }
  ],
  "architecture": "ChorusStack with four swappable ports: LLM backend, checkpoint store (SQLite free; Postgres persistence license-gated), semantic cache, retrieval (PrismRAG plug-in). Optional PrismGuard make_guard_handler before RAG/LLM. Route Ledger records why the agent routed. Canonical order: PrismGuard → ChorusGraph (+ PrismRAG) → LLM → PrismShine.",
  "useCases": [
    "Legal, healthcare, and finance agents that need auditable routing on a customer database",
    "Teams that want one runtime instead of assembling cache, retrieval, memory, and tracing",
    "Repeat-intent workloads where a default-on semantic cache at routing boundaries matters"
  ],
  "benchmarks": {
    "published": true,
    "summary": "Published Azure scale run heavy_20260708_140300 (n=300/scenario, seed 42, Gemini): task success 4–15 percentage points, mean LLM calls 31–76% lower, mean latency 8–73% lower vs the documented LangGraph baseline.",
    "primaryRun": {
      "id": "heavy_20260708_140300",
      "label": "Azure scale run",
      "environment": "Azure ACI (16 GB)",
      "model": "Gemini",
      "seed": 42,
      "tasksPerScenario": 300,
      "summary": {
        "successPointGainRange": "4–15 percentage points",
        "meanLlmCallReductionRange": "31–76%",
        "meanLatencyReductionRange": "8–73%"
      },
      "scenarios": [
        {
          "name": "Finance single-agent",
          "langGraphSuccessPct": 90,
          "chorusGraphSuccessPct": 96.7,
          "langGraphMeanLlmCalls": 3.33,
          "chorusGraphMeanLlmCalls": 0.8,
          "langGraphMeanLatencyMs": 4972,
          "chorusGraphMeanLatencyMs": 1318
        },
        {
          "name": "Finance multi-agent",
          "langGraphSuccessPct": 89,
          "chorusGraphSuccessPct": 93,
          "langGraphMeanLlmCalls": 2.04,
          "chorusGraphMeanLlmCalls": 0.75,
          "langGraphMeanLatencyMs": 3081,
          "chorusGraphMeanLatencyMs": 1335
        },
        {
          "name": "Healthcare single-agent",
          "langGraphSuccessPct": 73.7,
          "chorusGraphSuccessPct": 84,
          "langGraphMeanLlmCalls": 2.94,
          "chorusGraphMeanLlmCalls": 1.33,
          "langGraphMeanLatencyMs": 7105,
          "chorusGraphMeanLatencyMs": 3812
        },
        {
          "name": "Healthcare multi-agent",
          "langGraphSuccessPct": 62.3,
          "chorusGraphSuccessPct": 77.3,
          "langGraphMeanLlmCalls": 3.85,
          "chorusGraphMeanLlmCalls": 2.67,
          "langGraphMeanLatencyMs": 10354,
          "chorusGraphMeanLatencyMs": 9537
        }
      ]
    },
    "disclosures": [
      "Integrated product stack comparison, not an engine-only microbenchmark.",
      "ChorusGraph includes its productized cache, memory, Route Ledger, and deterministic routing; the LangGraph baseline is a competent framework implementation without those ChorusGraph layers.",
      "No published performance claim against CrewAI.",
      "No published performance claim against Microsoft Agent Framework.",
      "No published performance claim against AutoGen."
    ],
    "urls": [
      "https://github.com/insightitsGit/ChorusGraph/blob/master/docs/BENCHMARK.md",
      "https://github.com/insightitsGit/ChorusGraph/blob/master/docs/BENCHMARK_RESULTS.md",
      "https://github.com/insightitsGit/ChorusGraph/tree/master/benchmark/results/azure_heavy_20260708_140300"
    ]
  },
  "github": "https://github.com/insightitsGit/ChorusGraph",
  "pypi": "https://pypi.org/project/chorusgraph/1.3.0/",
  "documentation": [
    {
      "title": "Python agent runtime guide",
      "url": "https://www.insightits.com/guides/python-agent-runtime.html"
    },
    {
      "title": "ChorusGraph vs LangGraph",
      "url": "https://www.insightits.com/compare/chorusgraph-vs-langgraph.html"
    },
    {
      "title": "Agent frameworks guide",
      "url": "https://www.insightits.com/compare/agent-frameworks.html"
    },
    {
      "title": "Whitepaper",
      "url": "https://github.com/insightitsGit/ChorusGraph/blob/master/docs/WHITEPAPER.md"
    },
    {
      "title": "Benchmark results",
      "url": "https://github.com/insightitsGit/ChorusGraph/blob/master/docs/BENCHMARK_RESULTS.md"
    },
    {
      "title": "Benchmark methodology",
      "url": "https://github.com/insightitsGit/ChorusGraph/blob/master/docs/BENCHMARK.md"
    },
    {
      "title": "interactiveDemo",
      "url": "https://insightitsgit.github.io/ChorusGraph/demo.html"
    },
    {
      "title": "fairness",
      "url": "https://github.com/insightitsGit/ChorusGraph/blob/master/benchmark/FAIRNESS_H9.md"
    },
    {
      "title": "scaleRun",
      "url": "https://github.com/insightitsGit/ChorusGraph/tree/master/benchmark/results/azure_heavy_20260708_140300"
    },
    {
      "title": "installGuide",
      "url": "https://github.com/insightitsGit/ChorusGraph/blob/master/docs/INSTALL.md"
    },
    {
      "title": "aiIdePrompts",
      "url": "https://github.com/insightitsGit/ChorusGraph/blob/master/docs/AI_IDE_PROMPTS.md"
    },
    {
      "title": "plugins",
      "url": "https://github.com/insightitsGit/ChorusGraph/blob/master/docs/PLUGINS.md"
    }
  ],
  "install": "pip install \"chorusgraph==1.3.0\". Apache 2.0. Published Azure scale benchmark: +4–15 pp task success vs the documented LangGraph baseline (n=300/scenario).",
  "version": "1.3.0",
  "relatedProducts": [
    "prismguard",
    "prismrag",
    "prismshine",
    "choruscontrol"
  ],
  "publisher": {
    "@type": "Organization",
    "name": "Insight IT Solutions",
    "legalName": "Insight IT Solutions LLC",
    "url": "https://www.insightits.com"
  }
}