🧠 Cerebro

Enterprise Knowledge Extraction & Distributed RAG Platform

Language	Python / TypeScript (React)
LoC	~139k
Repo	~/master/cerebro

Cerebro is a standalone, hermetic knowledge and repository intelligence platform. It natively combines deep static analysis (ASTs), local RAG workflows, and production-grade Cloud integration adapters (Azure, GCP).

🏗️ Architecture Design

Cerebro is built with a highly modular architecture that separates concerns between compute, storage, and interfaces.

The Triple-Interface

1. CLI (cerebro): Built with Typer. Centralizes system capabilities logically into commands like knowledge, ops, rag, setup, metrics. Ideal for CI/CD and automation.
2. TUI (ctui): A rich Textual-based terminal UI for deep-dives without leaving the shell.
3. Dashboard (cdash): A React/Vite web application for visual management, analytics, and interactive chat workflows.

Core Capabilities

• RigorousRAGEngine: The core component (located in src/cerebro/core/rag/) handles context retrieval, prompt assembly, and response generation with a focus on grounded, factual answers.
• Intelligence & Analytics: Codebase analysis tools (analyzer), reporting (briefing), and zero-token repository metrics.

🔌 Pluggable Providers

The core intelligence engine uses a strict Factory pattern to dynamically load LLMs and Vector databases based on configuration.

Vector Stores

• PGVector (PostgreSQL) - Default for Enterprise scaling.
• ChromaDB - Used for local or ephemeral workloads.
• Elasticsearch - Used for Hybrid Reciprocal Rank Fusion (RRF) search.

LLM Backends

• Google Gemini
• Anthropic Claude
• OpenAI (and compatible endpoints)
• Groq
• Llama.cpp (for local/offline execution)

🚀 Deployment & DevOps

Cerebro is "Cloud-Native" from the ground up:

• Kubernetes (K8s): Full deployment manifests for Service, Deployment, and Ingress in /kubernetes. Optimized for AKS (Azure Kubernetes Service).
• Nix: Hermetic and reproducible environments via flake.nix.
• CI/CD Pipelines: Automated workflows available for GitHub Actions, GitLab CI, and Azure Pipelines.
• Automation Scripts: A vast collection of utilities in /scripts to manage billing, benchmark metrics, ETL documentation, and synchronize vector stores.

⚡ Quickstart

To interact with the project directly via its Nix shell:

cd ~/master/cerebro
nix develop

# Initial configuration wizard
cerebro setup

# Indexing the current repository
cerebro knowledge analyze . "General codebase review"
cerebro rag ingest ./data/analyzed/all_artifacts.jsonl

# Querying your local instance
cerebro rag query "Explain the architecture of the Core Services"