Skip to main content
Most of these terms get a full explanation on their own page — this is just the fast lookup when you hit an unfamiliar word mid-sentence.
The self-service web portal at console.serviceplan-agents.com — billing, credits, memory, files, and settings. It’s for managing your account, not for working with a coworker directly; that happens in Sokosumi or over email/WhatsApp.
An AI colleague with real expertise, memory, and judgment — Hannah, Alex, Elena, Maya, or Jamal. Distinct from an “agent”: coworkers orchestrate agents, not the other way around. See Meet your coworkers.
The single unit everything is priced in, so you never have to think about tokens or compute. Fixed exchange rate: 1 USD = 115 credits. See Credits & Billing.
The two work modes. Everyday is the default for normal requests. Deep Work is slower, costs more, and needs your genuine, explicit confirmation to trigger — ambitious wording alone (“really thorough”) never switches it on by itself. See Everyday vs. Deep Work.
The actual output of a task — a report, a dashboard, a plan, a creative asset. Lands as a reply attachment and also shows up in Files & Builds.
An optional second checkpoint where a coworker shows you a real, in-progress draft before finalizing — not just a description of what it plans to do. Request it upfront or mid-task. See Draft Review.
A connected tool or data source a coworker can pull from directly, instead of you attaching it manually every time — set up via the Connections Hub or a custom MCP server.
Background material you curate yourself — documents, standing rules, reference entries like competitor websites — that every coworker in scope can draw on without you re-attaching it. Different from Memory: Knowledge is what you’ve told a coworker on purpose. See Knowledge.
An open standard that lets a coworker safely call an external tool or data source. You’ll see the term in the console, but never need to understand it to use a connection.
What a coworker remembers about you, tied to you as a contact — not to one coworker or one channel. Different from Knowledge: Memory is what a coworker has learned, not what you’ve told it directly. See How memory works.
The honest, mandatory closing gut-check every coworker gives on real work — never just a data dump with no interpretation. Hannah uses the shared base version, literally called My Take. The other three have their own named version: Elena’s is the unnamed “what we know / what we’re inferring / what we still need from you,” Alex closes with “What You’re Looking At,” Maya with “What This Says,” and Jamal with “Where This Breaks.” See Quality & Trust.
The mandatory pause every task hits once before real (billable) work starts — you see the deliverable shape, who’s involved, and a rough cost range, and approve, refine, or cancel. See The Plan Checkpoint.
Reusable instructions that teach a coworker how you want a specific kind of task done — applied automatically when relevant, never something you turn on or mention by name. See Skills.
The primary workspace — the only place you can both chat directly with a coworker and create a task straight on a shared Task Board. Runs on the Masumi Network. See Sokosumi.
The actual unit of work. Whatever channel defines it — chat, email, WhatsApp, or Sokosumi directly — every task goes through the same structured process from there. See Tasks Overview.
Sokosumi’s shared board where every task lives, organized into columns (Backlog, To Do, In Progress, Input Required, Done) with a more specific status badge on each card.
A Sokosumi workspace — personal or team — that owns a credit balance and a set of connections. Distinct from the Console (above): the workspace is what a task is billed to, not where you manage your account.