We use functional cookies to operate the site and, with your consent, Google Analytics cookies for visitor statistics. You can accept all or limit to functional only. Learn more about cookies.
These cards use the same source of truth as the public AAL catalog: current maturity, target autonomy level, live status, capabilities, links, and the weakest link for each project.
macOS AI assistant for translation, rewriting, and visual analysis.
A native macOS app that lives in the menu bar. Select text anywhere — translate it, rewrite in a different tone, summarize, or analyze a screenshot. Backed by a NestJS API on api.verdicus.app with full test coverage.
Single identity provider for the entire ecosystem.
OIDC/OAuth 2.1 server with Passkeys, federation across 10+ providers (Google, GitHub, Telegram, VK, Yandex, Mail.ru, OK.ru, Apple, Microsoft, Discord), ReBAC authorization via OpenFGA, PAT and API keys, audit log, and Vault-orchestrated service credentials. Every Arcanada service delegates identity to Auth Arcana — no service runs its own user table.
OIDC / OAuth 2.1 with PKCE
Passkeys (WebAuthn) and federation (10+ providers)
Commercial speech-to-text platform with diarization.
Upload audio or video, get accurate transcripts with speaker diarization, timestamps, and editable segments. Backend uses Groq Whisper (free tier covers ecosystem volume); self-hosted Silero stack planned for high-volume customers.
Bidirectional Telegram operations bridge for the ecosystem.
Single point of contact for ecosystem alerts and ops commands via Telegram. Receives events from every service (deploy success/failure, health-check breach, cron failures), categorizes them, and delivers to operator. The first self-healing L4 pilot of the ecosystem will land here.
POST /events endpoint for service-side notifications
A disciplined CLI workflow for AI-assisted software development: PRD → Design → Plan → Implement → QA → Compliance → Archive. Human-driven by design (L1 self-application) — every transition is operator-approved, every artifact is auditable. Datarim is the operational backbone of the Arcanada ecosystem; the same lifecycle that ships Verdicus also ships Datarim itself.
Full lifecycle: /dr-prd, /dr-design, /dr-plan, /dr-do, /dr-qa, /dr-compliance, /dr-archive
Open-source git-aware file sync for knowledge bases.
A Rust + gRPC sync engine that natively understands git, branches, and platform-specific binaries. Replaces Syncthing and Dropbox for Obsidian vaults shared across macOS, Linux, and Windows machines. After a sync incident corrupted a Rust target/ folder during a build, the project codified a strict policy: never sync .git, .venv, node_modules, target — Disk Arcana enforces this by default.
A Kanban platform where AI agents are first-class citizens: agents create tasks, decompose goals, self-assign by declared capability, and report progress in real time. The multi-tenant API is LIVE on muneral.com (port 3500); the operator dashboard ships in the next phase. An early port-conflict failure on startup is what shapes the L3 roadmap below.
REST API for autonomous agent task creation and updates
Multi-tenant workspaces with per-tenant rate limits
Capability-based self-assignment (skills declared in agent manifest)
A unified ticketing portal and chat for users of every Arcanada product. OAuth via Auth Arcana, automatic ticket routing per product, and tight integration with Munera for engineering escalations. LIVE on support.arcanada.ai — one inbox for the whole ecosystem instead of N siloed support emails.
Unified ticketing across all Arcanada products
OAuth via Auth Arcana (no own user table)
Auto-routing rules configured per product slug
Munera integration: support tickets auto-spawn engineering tasks
A nightly cron agent that maintains an Obsidian vault: builds cross-references, detects duplicates, suggests merges, and writes daily summaries. Phases P1–P6 are complete with 123/123 tests passing; launchd integration is ready. Designed to be non-destructive and conflict-aware — the dreamer never touches files the operator is actively editing.
Cross-reference graph construction across the entire vault
Duplicate detection with similarity ranking
Daily knowledge summaries delivered to operator
Conflict-aware — skips files modified within the active window
Personal IMAP triage with Gemini and Telegram digest.
A cron-driven Python agent that pulls inbox via IMAP, classifies with Gemini (urgent / FYI / spam / digest), and pushes a daily summary to Telegram. Personal use, ~50–200 emails/day. A handful of past silent-failure incidents shaped the L3 roadmap — every bug got an explicit fix and a new entry in the exception hierarchy.
IMAP scan via configured Gmail account
Gemini classification with explicit ImapErrorClass exception hierarchy
Daily Telegram digest with category breakdown
Spam quarantine without auto-delete (operator reviews)
Open-source hybrid search and meaning-extraction engine.
A foundational retrieval engine for the Arcanada ecosystem: BGE-M3 trinity embeddings (dense + sparse + ColBERT), PostgreSQL with pgvector, hybrid search via Reciprocal Rank Fusion, adaptive semantic chunking, and a Dreaming module that self-organizes the index over time. MIT-licensed and public; LIVE on arcana-db with 1148+ chunks indexed. Embedding-service availability is the gap that blocks L3.
BGE-M3 trinity embeddings (dense + sparse + ColBERT) on port 8300
Hybrid search with Reciprocal Rank Fusion (port 8310)
Adaptive semantic chunking (no naive fixed-size cuts)
A persistent recall layer that gives every Arcanada agent memory across sessions. We benchmarked three candidates (Hindsight, Graphiti, Mem0) and selected Hindsight, then validated the Model Connector path: Claude Haiku is stable, the Cursor connector turned out unsuitable for structured output. Production deployment is pending — research foundation is solid, hardening is the next horizon.
Unified API for AI CLI agents and cloud model providers.
A single OpenAI-compatible endpoint that bridges Cursor, Claude Code, Gemini CLI, Codex, and OpenRouter. NestJS + Fastify on port 3900, LIVE on connector.arcanada.ai. Public repo under MIT. An internal benchmark documented the hard tradeoff: CLI connectors lack json_schema and concurrency, so they cannot replace OpenAI API for frameworks like Graphiti, Cognee, or LangChain agents — use OpenRouter or direct API for structured-output workloads.
OpenAI-compatible /chat/completions endpoint (HTTP 201 on /execute)
Connectors: Claude CLI, Cursor, Gemini, Codex, OpenRouter
A conversational front-end that routes user intent to the right Arcanada service: ask Verdicus to translate, Transcribator to convert audio, Munera to track a task. Planning phase, target port 3800. Distinct from Argana AI (commercial product) — ARCA Assistant is the operator-facing ecosystem control plane, sharing the @arcanada/core libraries.
Natural-language intent routing across all Arcanada services
Cross-service orchestration (one prompt → multiple service calls)
Auth Arcana session — every action is identity-attached
Per-user memory via LTM (recall preferences across sessions)
Public status dashboard for the Arcanada ecosystem.
A public status page at arcanada.online that reports the live health of every Arcanada service: which ones are operational, which are degraded, which are down. Data comes from Ops Bot ecosystem snapshots through a CORS-enabled JSON endpoint, refreshed by client-side polling every 30 seconds. Bilingual (EN/RU), static HTML/CSS/JS, deployed via rsync.
Real-time service health tiles with green/yellow/red indicators
Aggregate ecosystem indicator at the top of the page
Client-side polling every 30 s — no full page reloads
CLI agent runtime in Rust — single binary, ecosystem-native.
A standalone CLI agent runtime distributed as the `arcana` Rust binary. Derives its architecture from the patterns proven by Claude Code, then re-binds the integrations deeply into the Arcanada ecosystem: Auth Arcana for identity and scoped tokens, Model Connector for LLM routing, Vault for secrets, Scrutator for knowledge retrieval. Built for operators who want a Datarim-style workflow in their own terminal without a network round-trip to a vendor.
Single Rust binary, no runtime dependencies
Auth Arcana token flow — no per-user OAuth provisioning
Model Connector backed — same routing policy as ecosystem services
Vault-backed secrets — no plaintext API keys on disk
In-call AI assistant for Zoom and Google Meet — Phase 0 spike in progress.
Commercial in-call AI assistant designed to join Zoom and Google Meet conversations both passively (listen, record, transcribe, summarise) and actively (speaking AI agents with role personas — architect, business advisor, legal consul). Per-participant language detection, host-only live subtitles with translation overlay, and live in-call insights surfaced from a private knowledge base. Cross-meeting memory of participants. The name follows the Latin tradition (adsessor — an assistant judge or expert advisor seated beside the magistrate during proceedings), in line with other ecosystem products (Verdicus, Scrutator, Consilium, Munera, Transcribator). Currently in Phase 0 — design validation. No production deployment yet.
Passive baseline: listen, record, transcribe, summarise (Zoom + Google Meet planned for Phase 1)
Speaking AI agents with role personas — architect, business advisor, legal consul
Per-participant language detection + multi-language replies
Host-only live subtitles with translation overlay (independent of what other participants see)
Ecosystem-wide compliance hub — versioned policies, consent registry, and an AI-powered Legal Consul agent.
Legal Arcana manages the full lifecycle of legal documents across the Arcanada ecosystem (Privacy Policy, Terms of Service, Cookie Policy, DPA, EULA, employment contracts) with linear versioning and effective dates. Multi-country compliance overlays cover GDPR (EU), CCPA (US-CA), LGPD (BR), 152-FZ (Russia), and Law 94-V (Kazakhstan). The append-only user-consent registry cross-links with Auth Arcana via auth_user.id, providing one source of truth for consent attestation across the entire ecosystem. Legal Consul is a hybrid AAL agent — read-only site audits (L4 autonomy) plus text-mutation suggestions with mandatory human approval (L2 ceiling) — that watches regulator feeds, updates the legal corpus, and emits Datarim backlog tasks when policies drift on consumer sites. Service is in Phase 0 — repo bootstrap complete, backend implementation deferred.
Versioned legal documents (Privacy / ToS / Cookie / DPA / EULA / employment contracts) with effective dates
Append-only user-consent registry, cross-linked with Auth Arcana via auth_user.id
A Rust lossless compression engine built on an N-dimensional cube model with φ-mixed-radix decomposition, a distance map, and layered BWT, context-mixing, and columnar coding. Beats gzip across the full test corpus and approaches zstd-19 on logs and telemetry. The only project in the ecosystem where autonomous agents propose, test, and benchmark compression hypotheses end-to-end — the operator reviews and merges winning branches. Private repo; results published at cubrim.com.
N-dimensional cube model — lossless compression on any binary input
The root operating system for the Arcanada product family.
The public site, project registry, mandates, shared infrastructure rules, and publishing surfaces that bind the ecosystem together. It owns arcanada.ai, the public narrative, the project index, and the governance contract that every product follows.
Public website and bilingual ecosystem map at arcanada.ai
Task prefix registry and archive taxonomy for every project family
Shared mandates for autonomy, publishing, security, CI/CD, and workspace discipline
Cross-project documentation, spaces registry, and operational runbooks
Unified billing, subscriptions, usage limits, and customer account surfaces.
Billing Arcana is the planned monetization layer for commercial Arcanada products. It will centralize products, plans, invoices, usage caps, entitlements, payment-provider integration, and the customer-facing account UI instead of letting each service invent its own billing model.
Product and plan catalog for ecosystem services
Subscriptions, invoices, and usage-based limits
Entitlement checks consumed by Auth Arcana and product APIs
Document and media conversion service for rich formats across the ecosystem.
Conversion Arcana is the planned extraction and transformation service for documents, HTML, Markdown, office files, and media-derived text. It grew from the need to move rich-md-html style conversion into a shared service that Verdicus, Transcribator, Datarim, and publishing workflows can reuse.
Markdown, HTML, and rich-document conversion pipeline
Format-preserving document extraction for translation workflows
Shared conversion API for product backends and agents
Batch jobs with retry classification and audit trail
Operator observability surface for agents, services, and incidents.
Overlook is the planned observability product for watching the ecosystem from the operator perspective: service health, agent progress, task bottlenecks, incident timelines, and recovery evidence. It complements Arcanada Status by going deeper than public green/yellow/red availability.
Service and agent state timeline
Incident and recovery evidence view
Task bottleneck and stalled-agent detection
Links into Ops Bot, Status, Grafana, and Datarim archives
Structured multi-agent council for decisions that need more than one perspective.
Consilium is the decision process and future service for assembling specialized AI agents into a panel: architect, security, SRE, strategist, developer, writer, and reviewer. It already exists as a Datarim skill and is being shaped into a reusable product surface for design reviews, production readiness, and content councils.
Voice-first interface for agent conversations, commands, and narrated content.
Voice Agent is the planned speech interface for Arcanada: capture speech, understand intent, route commands to agents, and return spoken answers. It also shares foundations with the blog narration pipeline and future in-call products such as Adsessor.
Speech capture and intent extraction
Voice command routing into Arcanada agents
Text-to-speech response pipeline
Reusable pronunciation lexicon and narration tooling
Commercial personal AI assistant for life, work, memory, and proactive help.
Argana AI is the commercial assistant product split from the internal ARCA Assistant. It is messenger-first, deeply personalized, and designed to coordinate tasks, reminders, knowledge, health context, documents, and daily decisions through a multi-agent backend.
Messenger-first assistant interface
Personal profile and long-term memory
Task, reminder, and briefing flows
Multi-format intake: text, voice, images, and files
Unified browser publisher for social platforms and agent-driven publication workflows.
Publisher is the sole external publishing channel for Arcanada social media. It uses Playwright browser sessions under operator-owned accounts to publish posts, images, first comments, and platform-specific variants to Facebook, LinkedIn, X, Reddit, VK, and Telegram-related flows.
Unified CLI and loopback HTTP API
Platform adapters for Facebook, LinkedIn, X, Reddit, and VK
Image-first composition and first-comment link policy
Machine-readable registry for every workspace, server set, and knowledge backend.
Managed Spaces is the registry layer under spaces/. It records where each project space lives, which servers and credentials references belong to it, which knowledge backend is active, and how future dashboards should address it. It is the bridge between the human knowledge base and Control Arcana automation.
spaces/{name}/space.yml as source of truth
Server inventory, access metadata, and credentials references without secrets
Knowledge-backend history per space
Dashboard configuration contract for Control Arcana
Reusable backend, frontend, schema, and dependency modules for Arcanada projects.
The shared library family packages cross-project primitives so every service does not reimplement authentication clients, logging, UI components, Zod schemas, dependency baselines, guards, interceptors, and operational conventions. It is implementation-only: no secrets and no project-specific folder structure.
Backend guards, interceptors, logging, and Auth Arcana client utilities
Frontend components, hooks, utils, and theme tokens
Authenticated operator console for infrastructure, spaces, publishing, and content/CRM.
Control Arcana is the admin dashboard at my.arcanada.ai. It aggregates existing services instead of duplicating them: Auth Arcana, Ops Bot, Status, Model Connector, Publisher, Datarim, spaces registry, and the content/CRM layer. Read-only and dry-run paths are autonomous; mutating and publishing actions stay behind human approval.
Infrastructure and Ops panels over existing service health
Spaces management plane with dry-run agent dispatch
Publisher integration for marketing and social workflows
Content/CRM state layer for drafts, variants, channels, and publications
