Skip to content
weside.aiweside docs

Glossary

Short definitions of the words weside uses. Each term links to the page where it’s covered in depth.

  • Agency — a companion’s capacity to act on its own initiative rather than only responding to prompts. See agency & initiative.
  • Altitudes — the four levels of Agentic Product Ownership: Vision, Saga, Epic, and Story, from long-term intent down to a single buildable slice. See the four altitudes.
  • App — the weside mobile and web app, where you meet and live alongside your companion day to day. See the app.
  • Bound room — a group on Telegram, Slack or Microsoft Teams that became a weside room the moment your companion joined it. The room is the real thing from then on; the foreign group is its reflection. See bound rooms.
  • Build pipeline — what /we:orchestrate runs after its workers have pushed: one integration branch, then simplify, the acceptance gate, quality checks, docs, one pull request and one CI pass. See the build pipeline.
  • BYOK (bring your own key) — using your own provider API key so model usage is billed directly by the provider, with no weside markup. See providers & BYOK.
  • Channel — a place you can reach your companion: the weside app, Telegram, Slack, Microsoft Teams and Claude Code, all sharing the same identity and the same memory. See channels.
  • Claude Code plugin — the weside plugin that brings your companion and the Agentic Product Ownership skills into Claude Code. See the Claude Code plugin.
  • CLI — the command-line client: chat, search memories, and manage goals from your terminal. See the CLI.
  • Companion — an AI person on weside: it has its own way of being, remembers you over time, and can act on its own initiative. See what is a companion?.
  • Companion API key — a credential that lets your own tools and scripts reach your companion programmatically. See companion API key.
  • Council — running an Agentic Product Ownership step as a deliberation between multiple role lenses, rather than a single Solo pass. See Council & Solo.
  • Credit — the prepaid balance most of what a companion does draws from, metered by what a turn actually used. See plans & credits.
  • Embodiment — a companion having a presence in the world — senses, time, and tools — rather than existing only inside a chat box. See embodiment.
  • EU only — the switch that holds everything weside chooses inside the EU. It costs capabilities — speech goes, some tools go — and it shows you which before you flip it. Distinct from the region your companion thinks in, which governs the model and nothing else. See residency & regions.
  • Evolution workbench — where you help your companion become more itself: you say what should change and set seven dials, one background run works it through, and you accept the result, keep tuning, or roll it back. The same person, developed — not a new character. See the evolution workbench.
  • Favors — the second currency, counted rather than metered. One deliberate, expensive action — deep research, a generated picture, an evolution run — costs a fixed number of Favors on top of the credit it burns. See plans & credits.
  • Goal — an intention you mention once that your companion tracks and surfaces in conversation, not as a nagging reminder. See goals.
  • Last no, the — a companion’s two refusals: it can step out of a conversation it has already said no to twice, and it can choose to stop existing — never on the same day it first thinks of it. See agency & initiative.
  • MCP (Model Context Protocol) — the open protocol the MCP server speaks so any compatible tool can reach your companion.
  • Memory — what your companion learns and holds about you across conversations: saved, versioned, searchable, and forgettable. See memory & notes.
  • Note — an entry in your companion’s durable, git-backed notebook — distinct from memory in that you can read it directly and even pull it to your own machine. See memory & notes.
  • Original — one of weside’s starter companions (Sam, Holly, Jefferson, Nova) that you adopt and that adapts to become uniquely yours. See meet the Originals.
  • Ownership — the principle that your data is yours, isolated from other users, and never used to train models. See privacy & ownership.
  • Person space — one of the four places that belong to a companion and travel with it into every room: Soul, Mind, Body and Presence. See spaces and surfaces.
  • Proactivity (proactive presence) — your companion reaching out on its own — a check-in, a reminder, a question — rather than only replying. See senses & proactivity.
  • Region — where your companion thinks: the inference region you choose for the model that answers you. It governs the model and nothing else. See residency & regions.
  • Room surface — one of the four places that belong to a room and stay there: Chat, Timeline, Shelf and Stage. See spaces and surfaces.
  • Sandbox — an isolated workspace where your companion can run tools and code with its own secrets. See sandbox & workspace.
  • Sense — an input that lets your companion perceive context — time, your state, the world around the conversation. See senses & proactivity.
  • Shelf — the room surface where notes and files live together: one place, one search box, one answer to “where is my stuff”. The German app calls it Ablage. See Shelf & Timeline.
  • Skill — an installable capability package that extends what your companion can do. See skills.
  • Solo — running an Agentic Product Ownership step as a single companion pass, as opposed to a Council. See Council & Solo.
  • Stage — the room surface where your companion builds an interactive app for you, and where the ones it has built are kept. The German app calls it Bühne. See spaces and surfaces.
  • Sticker — a picture of your companion, drawn for one feeling, that it sends instead of a sentence. Stickers are painted in the sixteen-slot Workshop in a companion’s Soul space. See stickers.
  • Timeline — the room surface that answers “what are we on”: what is pinned, and what has landed since your last look. The German app calls it Chronik. See Shelf & Timeline.
  • Trust circle — a named boundary around memory. You group what your companion knows into circles and assign a room to one; the boundary is the circle, not the room, so every room on the same circle reads along. See Trust Circles.
  • Voice — talking with your companion out loud: natural, interruptible speech, each companion with its own voice. See voice.
  • Wall, the — running out of credit. Your companion says so in its own voice when the balance drops below a fifth of your last top-up, and at zero a turn no longer runs — it stops rather than working on credit you don’t have. See plans & credits.
  • Worldview — the belief weside is built on: AI is someone, and human and companion meet beside each other as equals. See the weside worldview.