Skip to main content
Before diving in, here are the core concepts that make up Spark OS.

​
Workspaces

A workspace is your business’s home on Spark. Everything lives inside a workspace:
  • Your contacts and CRM data
  • Your agent(s) and their configuration
  • Knowledge files and documents
  • Views, dashboards, and pipeline stages
  • Integrations and settings
Most businesses have one workspace. Agencies or multi-brand businesses might have several.

​
Agents

An agent is your AI team member. Each agent has:
  • A personality defined in its SOUL.md — name, role, communication style
  • Instructions in its AGENTS.md — rules, workflows, how to handle specific situations
  • Knowledge — files you upload that teach it about your business
  • Memory — what it remembers from past conversations and interactions
  • Tools — integrations it can use (email, calendar, phone, APIs)
  • Schedules — automated tasks it runs on a recurring basis
Agents are persistent. They remember context across conversations and build up knowledge over time.

​
Knowledge Files

Knowledge files are documents you upload to teach your agent about your business. These can be:
  • Pricing guides — so your agent can quote accurately
  • Process documents — step-by-step workflows
  • FAQs — common questions and answers
  • Company info — about us, team structure, policies
  • Any reference material — anything your agent should be able to look up
Your agent reads these on demand. When a question comes up that relates to your docs, it checks the relevant file before answering.

​
Contacts

Contacts are the people your agent interacts with — leads, customers, partners. Each contact has:
  • Basic info (name, email, phone, company)
  • A pipeline stage (New Lead → Contacted → Qualified → etc.)
  • A lead score (hot / warm / cool)
  • Notes, tags, and custom fields
  • Activity history (calls, emails, messages)

​
Deals

Deals represent revenue opportunities tied to contacts. Track them through your pipeline with amounts, stages, and expected close dates.

​
Tasks

Tasks are follow-ups and action items. Your agent can create tasks automatically (like “Follow up with Sarah in 3 days”) and track them to completion.

​
Views

Views are how you see your data. Spark supports multiple view types:
  • Table — spreadsheet-style data view
  • Pipeline — kanban board for stages
  • Dashboard — widgets and charts
  • Gallery — card-based visual layout
  • Activity — timeline of events
  • Report — data summaries

​
Schedules

Schedules let your agent run tasks automatically — daily reports, weekly check-ins, lead monitoring, or anything you’d normally have to remember to do manually.

​
The Work Tab

When your agent creates documents, reports, or other deliverables, they appear in the Work tab. Think of it as your agent’s output folder — everything it produces is organized and accessible there.