Architecture

ContextUnity is a service mesh of modular Python services communicating over gRPC. Each service has a single responsibility and shares types through ContextCore.

Ecosystem Map

Communication Pattern

All inter-service communication uses the ContextUnit protocol. Every RPC call carries a typed Pydantic envelope with:

• Payload — domain-specific data (google.protobuf.Struct)
• Provenance — ordered trace of processing stages
• Security — scopes for access control
• Tracing — trace_id and parent_unit_id for distributed observability

Service Roles

Service	Role	Protocol	License
ContextCore	Shared kernel — types, protos, tokens, config	dependency	Apache 2.0
ContextRouter	Agent orchestration — LangGraph, tool dispatch	gRPC server	Apache 2.0
ContextBrain	Knowledge — vectors, memory, taxonomy, RAG	gRPC server	Apache 2.0
ContextWorker	Workflows — Temporal, schedules, agents	gRPC server	Apache 2.0
ContextShield	Security — token signing, prompt filtering	gRPC server	Commercial
ContextZero	Privacy — PII masking, persona engine	middleware	Commercial
ContextCommerce	E-commerce — PIM, product matching, enrichment	Django app	Commercial
ContextView	Observability — dashboard, traces, admin	Django app	Commercial

Technology Stack

• Language: Python 3.13+
• Inter-service: gRPC with Protocol Buffers
• Agent Framework: LangGraph (LangChain)
• Database: PostgreSQL 16+ with pgvector and ltree extensions
• Task Engine: Temporal.io
• LLM Providers: OpenAI, Anthropic Claude, Google Vertex AI, Groq, Perplexity
• Package Management: uv
• Configuration: Pydantic BaseModel with load_shared_config_from_env()