Arcanum

The Complete Domain Intelligence Suite

Seven projects unified under a single install — diagnostics, reference data, AI integration, agent skills, expert training data, a fine-tuned domain expert model, and conversational intelligence for the domain name industry.

Explore the Suite GitHub

The Suite

Seven projects, each with a distinct purpose, working in concert.

Seer

Domain Diagnostics

Multi-interface domain intelligence toolkit providing WHOIS and RDAP lookups, DNS resolution and propagation monitoring, SSL and DNSSEC analysis, availability checking, subdomain enumeration, and high-concurrency bulk operations.

20 Python functionsAvailability checkingDNSSEC validationSSL chain analysisCT log enumerationBulk ops (50 concurrent)
CLIRust libraryPython libraryREST APIMCP server
Repo
RustPyO3TokioFastAPI

Tome

Domain Knowledge Base

Reference database for internet TLDs, DNS record types with RFC references and examples, and a glossary of domain industry terminology. Data models complete, datasets not yet populated.

TLD databaseDNS record definitionsRFC referencesIndustry glossaryCross-database searchMultiple output formats
CLIRust libraryPython libraryREST APIMCP server
Repo
RustSQLitePyO3FastAPI

Tower

Unified MCP Server

Consolidates Seer and Tome into a single Model Context Protocol server exposing 22 tools through one entry point for AI assistants like Claude.

22 unified tools13 from Seer + 9 from Tomestdio transportClaude Desktop supportClaude Code supportModular dispatch
MCP server
Repo
PythonMCP SDKHatchling

Scrolls

AI Agent Skills

Skill definitions and reference documentation that give AI agents deep contextual knowledge about domain intelligence tools. Automatically bundled with Tower.

Skill definitionsCLI referencesAPI documentationPython binding guidesConfiguration examplesProgrammatic access
Python library
Repo
PythonMarkdownHatchling

Familiar

Conversational AI Agent

Natural language interface to domain intelligence built on LangGraph Deep Agents. 60 tools for diagnostics, security audits, penetration testing, brand protection, portfolio management, and persistent memory.

60 integrated toolsSecurity auditsPenetration testingBrand protection scansMulti-LLM supportSession memory
CLI
Repo
PythonLangGraphDeep Agents

Oracle

Training Dataset Generator

Generates expert-grade training datasets across 423 topics and 26 categories to fine-tune LLMs into domain name industry specialists. Combines synthetic generation, RFC extraction, and data augmentation with built-in quality validation.

423 curated topicsRFC extractionTool-use trainingMulti-format exportQuality validationDifficulty scaling
CLI
Repo
PythonPydantictiktokenClick
Coming Soon

Vessel

Domain Expert LLM

Fine-tuned language model trained on Oracle's dataset to become a domain name industry specialist. Built on Nemotron with reasoning traces, deployable locally via Ollama as a purpose-built domain intelligence engine.

NemotronDomain industry expertiseTool-use trainedReasoning tracesLocal deploymentOllama compatible
Model weightsOllama
Repo
UnslothLoRANemotron

Architecture

Layered by purpose — from raw diagnostics to natural conversation.

Interface Integration Core
Arcanum Meta-package
Familiar AI Agent
Tower MCP Server · 22 tools
Vessel Expert LLM
Oracle Training Data
Scrolls Agent Skills
Seer Rust core · 20 tools
Tome Rust core · 9 tools
Dependency
Uses / Bundles
1

Seer and Tome form the foundation — Rust performance cores with Python bindings providing raw domain diagnostics and reference data.

2

Tower consolidates both cores into a single MCP server. Scrolls provides contextual knowledge for AI agents. Oracle generates expert training datasets across 423 topics, which feed into Vessel — a fine-tuned domain expert LLM.

3

Familiar sits at the top — a conversational agent powered by Vessel, with 60 tools that leverage Seer, Tome, and Scrolls to answer questions in natural language.

Built With

Rust for speed. Python for reach. Protocols for intelligence.

Rust

Performance-critical cores compiled to native speed with zero-cost abstractions

Python

Orchestration layer with PyO3 bindings, FastAPI servers, and LangGraph agents

MCP

Model Context Protocol for seamless AI assistant integration via stdio transport

Tokio

Asynchronous runtime powering concurrent DNS queries and bulk operations

LangGraph

Deep Agents framework for planning, delegation, and long-conversation context

SQLite

Embedded reference database for TLD, record type, and terminology data

PyO3FastAPIPydanticHickory DNSReqwestSerdeClapHatchlingMCP SDKOllamaLangSmithtiktokenUnsloth