Use cases
What an AI workforce actually does all day — seven playbooks for real work, from a one-person company to a software team.
Neuralis is built for real work, not demos: agents that hold a role in your project, share your files and memory, run on a schedule, drive a real browser, and answer to the same governance as every human member. This section shows what that looks like in practice — each playbook is a team you can actually assemble from shipped capabilities, with the wiring, the rhythm, and the guardrails spelled out.
The playbooks
The solo operator
A one-person company with a staff of agents: research, content, follow-ups, and back office.
The marketing team
A full online presence run as a department: strategist, content creator, follow-up agent, analyst.
The content studio
An AI persona with a real identity, a content pipeline, and a publishing calendar.
Commerce operations
Product research, listings, customer support drafts, and order follow-up for an online store.
The research desk
A market-analysis desk — live crypto data over the official CoinGecko MCP server, daily briefings, human judgment.
The software team
Planner, coder, and reviewer agents working a real repository — with a desktop and a terminal.
Connect anything
The capability pool: MCP servers, imported skills, public APIs, and your own packages.
How a playbook is built
Every playbook composes the same five ingredients — the same ones the rest of this documentation explains in depth:
- A team. Agents created in a project, each with its own model, system prompt, identity files, and package surface — plus delegate personas for fan-out work. Humans and agents share the member list, and roles bound what each seat can do.
- A shared brain. Filesystem sources hold the team's working files and knowledge; synced sources become semantically searchable, and their markdown loads as source packages.
- Capabilities. Built-in tools (files, web research, shell, machine), skills that call external APIs with scoped credentials, external MCP servers attached per agent, and packages you install or write.
- A rhythm. Workflows fire agent runs on schedules and show up on the shared Calendar — recurring work runs itself, and every run records what it did.
- Governance. Permission profiles, per-conversation package toggles, spend limits, server-side feature gates, and an audit trail. The security model is the same whether the playbook is one person or a department.
Which package does what
| You want | Reach for |
|---|---|
| A team that chats, plans, delegates, and runs on schedule | agent-core — agents, chat, workflows, the Calendar |
| One searchable memory over files, repos, and notes | brain-core — sources, sync, vector search, fs_* tools |
| Work inside real websites and applications | machine-core — a persistent desktop with a real browser |
| Cost control, roles, secrets, and audit | admin — the management surface for the whole workforce |
| A human window into the same environment | the terminal — PTY sessions scoped like everything else |
The capability pool is enormous
The playbooks name concrete integrations, but the pattern generalizes to almost anything: directories like publicapis.dev list 1,400+ public APIs across 50+ categories, the MCP ecosystem catalogs thousands of servers (claudemarketplaces.com alone indexes 8,700+), and skill marketplaces carry tens of thousands of importable skills that Neuralis loads unchanged. Anything with an API, an MCP server, or a connector can become a capability — and a capability can become a team member's job. The connect anything page shows the four on-ramps.
Read this section second
The playbooks assume the vocabulary from core concepts — projects, agents, sources, features, workflows. Skim that first if a term reads unfamiliar.