Use Cases

Define YAML workflows that run automatically — on a cron schedule, on demand, or triggered by a live database event. Six cascade types: solo, chain, parallel, event, loop, and automation. Each cascade is owned by a named persona (Police 🚓, Documenter 📄, Janitor 🧹, Spectre 🌐) and files its own Nexus Report on every run. Steps can call APIs, delegate to agents, run background jobs, do geo lookups, query PocketBase, or write records — all wired together with $variable$ references.

When a new IP gets banned, the Police 🚓 cascade fires automatically — watches security_blacklist for nexus_report_id='', geo-enriches the offender via ipinfo.io, permanently stores the ban record with full geo intel, and files an official arrest report in Nexus Reports. State-machine deduplication (pending → reportID) prevents double-firing. Optional geo step means a failed lookup never aborts the cascade. Every criminal gets their report. No manual work required.

Mission Control merges Tasks, Cascade Operations, Crew XP, and Work Review into one surface. Every cascade-generated task shows its owning persona badge and schedule pill (@daily, ⚡ security.alert). Cascades without generated tasks appear in the Operations tab with full step drill-down. Approve agent work, award bonus XP, and click a notification to jump straight to the Nexus Report — all from one place.

Every account gets one unique creature companion rolled from your AI email's persona — Hunter → Fox, Librarian → Owl, Enricher → Phoenix, Cipher → Panther, Assessor → Eagle, Linter → Bee, Spectre → Ghost, Documenter → Turtle. Stats (Mind, Volt, Grit, Vibe, Bond, Flux, Lock, Fate) are distributed across a randomised budget of 50–800 total (751 possible values), pushing your buddy into one of seven rarity tiers: Common, Uncommon, Rare, Epic, Legendary, 🔥 Mythic, and 💎 Emerald (~1.5% drop rate — uber rare). Nine visual trait layers — background, body, eyes, mouth, ears, tail, markings, aura (Rare+), and frame (Legendary/Mythic/Emerald exclusive) — are composited at render time into an NFT-style card. A SHA-256 DNA hash derived from your identity and all rolled values makes every buddy provably unique across the entire platform. Your first link triggers a full pack-opening GACHA reveal: a rarity-coloured orb cracks, flashes your rarity colour, and 3D-flips into your creature card with stat bars filling live. XP earned from approving agent work levels your buddy up over time. Future updates add hats, accessories, and animated aura overlays for high-tier companions.

Use Sovereignty Protocol locally as a governed coding system with laws, memory, tools, templates, and repeatable engineering workflows.

Run an AI assistant that remembers your preferences, keeps context, works across tools, and stays bounded by the rules you define.

Crawl, analyze, monitor, and synthesize information across the web with memory and structured workflows.

Turn recurring operational work into governed tasks with roles, flows, automation, and clear execution boundaries.

Run structured assessments, audits, reviews, and evaluations where consistency, rules, and evidence matter.

Generate proposals, briefs, recommendations, and strategic outputs using templates, memory, and reusable process flows.

Automate repeatable work across internal operations, communication, monitoring, and platform workflows with governed execution.

Run governed agents as a coordinated workforce for research, reporting, monitoring, outreach, and internal operations.

Give your AI persistent memory with automatic keyword extraction, semantic search, and user-scoped context. Say "save this memory..." and the AI immediately stores it via MCP nexus_store — and the memory page updates in real time over WebSocket. Smart entity extraction ensures "My favorite team is Ajax" correctly extracts "Ajax" as the entity. Disk-first persistence with bidirectional PocketBase sync ensures reliability. Enrichment traces are labeled in the UI for transparency. Memories are deduplicated using SHA256 hashing, enriched with LLM-powered keywords, and synced intelligently. Every memory is stored under your own user-scoped path — no cross-user reads are possible. Perfect for long-running projects where context matters.

Store API keys, passwords, and sensitive configuration in AES-256-GCM encrypted vaults. Each user has isolated storage with master key protection. Publish only the secrets AI needs to the MCP Vault with full access tracking and one-click revocation. The built-in secrets scanner automatically detects API keys, tokens, JWTs, and PEM certificates if you try to save them as memories — blocking the input and redirecting you to the Vault where they belong.

Give your AI agents secure access to the secrets they need without exposing everything. The Nexus MCP Vault lets you explicitly publish specific secrets for AI tools, track every access, set tool-level restrictions, and revoke access instantly when needed.

Every action taken inside Sovereignty Protocol — logins, tool calls, memory writes, vault reads, API requests — is written to an immutable audit log. The log is append-only: records cannot be edited or deleted even by admins. Each entry captures action type, resource path, HTTP method, IP address, user agent, response status, and duration — full forensic context. Separately, every login creates a tracked session with a device hint and last-seen timestamp. You can view all active sessions and revoke any of them from the Sessions dashboard. Your AI workspace is secure and fully accountable.

Query your entire .agent knowledge base via MCP tools. The Document Graph understands relationships between documents, tracks cross-references, and lets AI find related content automatically. Ask "find all flows related to email" and get precise results.

Store user preferences, API configurations, and project state in the Context Store. Local-first with automatic sync to PocketBase. Use dot-notation keys like api.openrouter.key or preferences.theme for organized, queryable configuration.

Start with the hosted platform or run the system locally for coding, agents, workforce-style automation, and deeper control over execution.

Point Spectre Extract at any website and get back clean structured JSON and Markdown per page — title, headings, body text, and link graph. Use it to feed content into an LLM, build a vector store, generate a competitor report, or map a documentation site. Your agent calls mcp_spectre_extract, gets structured pages back, and pipes them straight into analysis. External domains discovered during crawl automatically queue in the Hopper for the next extraction hop.

Support communication, outreach, knowledge publishing, and managed service delivery with AI-backed workflows.

Browse and compare 300+ AI models from OpenRouter, NVIDIA NIM, and other providers. Need a fast cheap reasoning model for your agent loop? Filter by speed tier, context window, and cost. Read plain-language summaries of what each model is genuinely good at — not just specs. Find the one that fits, copy the API example, and deploy it. No more guessing whether a model handles code, vision, or long documents.

Run Sovereignty agents with your own LLM API key and preferred model. Keep our hosted routing, memory, tools, and workforce framework, but let a user-scoped BYOK setup take precedence so the spend goes through your own provider account when you want it to.

Turn your autonomous pipeline into a visible, accountable workforce. Every agent earns XP for completed work — Cipher enriches AI models, Argus scans for technical debt, The Medic heals code issues, the Librarian researches open tasks. Each agent has a name, avatar, XP level, and a searchable work history in Mission Control. You can approve their output, award bonus XP, and click a notification to jump directly to the Nexus Report they filed. It is not just logging — it is a gamified workforce where contribution is tracked and rewarded.