Getting Started
Go from sign-up to your first AI-assisted output in under 5 minutes. This guide walks through every step.
Prerequisites
Before you begin, make sure you have the following ready. BOSS runs entirely in the browser -- there is nothing to install locally unless you want to use the MCP integration or REST API.
Modern browser
RequiredChrome, Edge, Firefox, or Safari (latest 2 versions). BOSS uses WebSockets for real-time agent responses.
Email or SSO account
RequiredSign up with any email, or use Google / Microsoft SSO for single sign-on. Enterprise customers can configure SAML.
API key (optional)
OptionalOnly needed if you plan to use the REST API or connect via MCP. Generate one after sign-up in Settings > API Keys.
Network access
OptionalEnsure cintri.co and *.cintri.co are not blocked by your firewall or proxy. WebSocket connections on port 443 must be allowed.
Sign up at cintri.co
Head to cintri.co and create your account. You can sign up with email, Google, or Microsoft SSO. The free tier includes 3 studios, 5 agents, and 1,000 AI requests per month -- enough to evaluate the platform with real work.
No credit card required. Your workspace is provisioned instantly.
What you get on the free tier:
Create your organization
After signing up, you will create your organization. This is the top-level container for everything in BOSS: studios, agents, documents, and team members. All billing, permissions, and compliance settings are managed at the organization level.
You can update all of these later from /admin/settings. The industry and size selections influence which studios and agent configurations are recommended during onboarding.
Pick your industry and get recommended studios
Based on your industry selection, BOSS recommends a starter set of studios. A law firm gets Legal Research, Client Management, and Compliance studios. An e-commerce company gets Product Marketing, Customer Support, and Inventory studios. A consultancy gets Strategy, Client Delivery, and Knowledge Management studios.
You can accept the recommendations, add more from the marketplace, or start with a blank workspace and build custom studios. There are 42+ templates across 10 industries.
Example: Technology company starter set
Code reviews, architecture, sprints
Roadmaps, specs, user stories
Campaigns, content, analytics
CI/CD, monitoring, incidents
Each studio comes pre-configured with the right resource primitives, block types, and agent assignments for its function. You can customize everything after setup.
Meet your AI team
Every studio is staffed with named AI agents. These are not anonymous chatbots. Each agent has a name, a role, a set of capabilities, and a trust score that evolves based on how well they perform.
Routes tasks, manages workflow, coordinates multi-agent work
Plans, analyzes, recommends, builds frameworks
Finds, compares, synthesizes, cites sources
Builds, debugs, deploys, reviews code
Generates content, copy, designs, presentations
Designs schemas, APIs, system diagrams
Crunches numbers, builds dashboards, forecasts
Drafts emails, reports, stakeholder updates
Agents are assigned based on studio type. You can reassign them, adjust their capabilities, or bring in specialist agents from the agent registry. See the full agent roster for all 35+ available agents.
Open your first studio
Click into any studio to land on its canvas. The canvas is your workspace: a flexible, block-based surface where you organize work, view documents, chat with agents, and track progress.
You will see a greeting from the studio's resident agent. They already know your org context, the studio's purpose, and what resources are available. They are ready to work.
The sidebar shows your 7 resource primitives: Docs, Tables, Files, Notes, RAG Search, API, and Systems. Each is ready to use out of the box.
What the canvas looks like:
- Top bar: Studio name, agent selector, search, settings
- Left panel: Resource primitives (Docs, Tables, Files, etc.)
- Center: Block-based canvas with drag-and-drop layout
- Right panel: Chat with your agent, activity feed
Start a conversation
Type a message in the chat panel. Your agent will respond with structured, actionable output -- not generic answers. Ask it to draft a document, analyze data, create a project plan, or research a topic.
Try these first messages:
Agents can create and modify blocks on the canvas, generate documents, query your knowledge base via RAG search, and trigger workflows across studios. Everything they do is logged in the audit trail.
The more you work with your agents, the smarter they get. BOSS compounds intelligence through a Train, Align, Evaluate loop that closes on real business outcomes.
Connect via MCP
Already using Claude Desktop, Claude Code, or another MCP client? You can connect directly to your BOSS workspace. Add this to your MCP configuration:
Replace sk_cintrico_YOUR_KEY with your API key from Settings → API Keys.
Once connected, you get access to 30 tools across studios, documents, agents, chat, search, and admin. Verify the connection:
Supported MCP clients: Claude Desktop, Claude Code, Cursor, Windsurf, and any application implementing the Model Context Protocol specification.
Generate your API key
API keys authenticate both MCP connections and direct REST API calls. Each key has configurable scopes so you control exactly what a client can access.
- 1Navigate to /admin/settings/api-keys
- 2Click "Generate New Key"
- 3Name the key (e.g., "Claude Desktop", "CI/CD Pipeline")
- 4Select scopes (studios, documents, agents, chat, admin)
- 5Copy the key immediately -- it will not be shown again
Using your API key
With curl:
With TypeScript SDK:
What happens after onboarding
Once your workspace is set up, your agents are not sitting idle. Here is what the platform is doing behind the scenes to prepare your environment:
Context initialization
Each studio builds an initial context model from your org name, industry, and any documents you have uploaded. Agents use this context to ground every response in your specific situation.
Service Engine activation
The 5 governance agents (Guardian, Budget, Compliance, Security, Optimization) begin monitoring all workspace activity. They enforce policies from the first request onward.
Knowledge graph seeding
As you upload documents, create notes, and interact with agents, BOSS builds a Graphiti knowledge graph that connects entities, relationships, and concepts across your workspace.
Trust scoring begins
Every agent starts with a baseline trust score. As you accept, reject, or refine their outputs, their scores evolve. Higher trust unlocks more autonomous capabilities.
Troubleshooting
Common issues and how to resolve them.
My MCP connection fails with "unauthorized"
Ensure your API key starts with sk_cintrico_ and has not been revoked. Keys are shown only once at creation -- if you lost it, generate a new one from Settings > API Keys. Also confirm the key has the correct scopes for the tools you are trying to use.
Agent responses are slow or timing out
Response time depends on the AI model being used and task complexity. Claude Opus tasks can take 15-30 seconds for complex analysis. Check the health dashboard at /admin/health for current system status. If you are on the free tier, requests may queue during peak hours.
I cannot see studios that another team member created
Studios have role-based access control. Ask the studio owner to add you as a member from the studio settings panel. Organization admins can see and manage all studios from /admin/studios.
My webhook events are not being delivered
Verify your webhook URL is publicly accessible and returns a 200 status within 5 seconds. BOSS retries failed deliveries 3 times with exponential backoff (5s, 30s, 5min). Check the webhook logs at /admin/webhooks for delivery status and error details.
How do I reset an agent's trust score?
Organization admins can reset trust scores from the agent profile page. Navigate to /admin/agents, select the agent, and click "Reset Trust Score." This resets the score to the default baseline (50) and clears the performance history. Use this sparingly -- trust scores exist to protect your workflow quality.