Skip to main content

Glossary

Core concepts​

Organization: Your company's space in Revelry Ai. An organization has its own projects, members, Knowledge, integrations, and settings.

Organization Details: Labeled facts about your organization, such as your address, key contacts, or standard terms. Organization Admins manage them on the Organization Details page. Chat can look them up, admins can ask in chat to add or update them, and Skills can insert them into their instructions.

Project: The workspace for a goal or ongoing effort. Everything in Revelry Ai happens inside a project, including chats, artifacts, and Workflow runs. See Projects.

Artifact: A document, report, piece of code, or other output created in Revelry Ai, from chat, a Skill, a Workflow, or the API. Every content change records a version you can view and restore. See Artifacts.

Artifact Type: The format and purpose of an artifact, such as PRD, Roadmap, or User Story. Type and category are set separately, so a PRD can live in the Product category or the Engineering category.

Category: A folder that organizes artifacts in a project. When category folders are turned on, a project with categories shows them as folders beside its artifact table, such as "Product," "Engineering," and "Design," and categories can nest up to three levels. Contact your Revelry Ai team to enable. A category's groups narrow who can see the artifacts in it.

Approval Gate: The prompt that appears before a tool that can change data runs, such as a create, update, or delete through an integration. You can choose Allow for this chat to let that tool run without asking for the rest of the chat, Allow Once, or Deny. Read-only tools run without asking.

Tools: The navigation section for Agents, Skills, and Workflows. It opens on an overview with a card for each, the most used tools, and recent activity, and each has its own tab.

Knowledge (navigation section): The navigation section for Documents, Collections, and Data Sources. It opens on an overview with a card for each, the most used documents and collections, and recent activity, and each has its own tab.

Chat History: The page that lists your own past conversations in the current organization, except archived ones. It's searchable and grouped into Today, This Week, and Older Chats.

Account menu: Your name at the bottom of the left navigation, with the organization switcher if you belong to more than one organization. Organization Admins also see the credit balance there. It leads to your profile and account settings, and for admins, the organization's settings.

Breadcrumbs: The trail at the top of each page that shows where you are, ending in the page title. On narrow screens a long trail shortens in the middle and can be expanded.

Ways to work​

Chat: The conversational interface. You explore ideas and create artifacts by talking with Revelry Ai, which calls tools as it needs them. See Chat overview.

Agent: A named assistant your organization configures with its own instructions and a curated set of Skills. Chatting with an Agent applies its instructions for the whole conversation. See Agents.

Workflow: A saved sequence of steps, each running a Skill, for work you repeat the same way every time. See Workflows overview.

API: Programmatic access for your own applications. The API can create artifacts, start and monitor Workflow runs, send chat messages, and manage projects and other resources. See the API Usage Guide.

Workflow components​

Step: One stage of a Workflow. Each step runs a Skill and can use outputs from earlier steps through links.

Link: A connection that routes one step's output into a later step's input.

Trigger: How a Workflow run starts. You can start a run from the app or through the Workflows API.

Workflow Run: One execution of a Workflow. A run tracks progress step by step, along with its status, the artifacts it created, and its processing logs. See Workflow Runs.

Validation step: A step you add to check the output of earlier steps for accuracy or completeness.

Prompts and Skills​

Prompt: The text you type to chat or an Agent in the moment, used once. See Prompts vs. Skills.

Skill: A saved, reusable instruction that runs as a self-contained task. A Skill combines instruction text with user inputs filled in at run time, document attachments, a model, built-in tools, and integrations. Skills were called Prompt Templates before v1.2.0. See Skills.

Skill Library: The Skills available in your organization, including Revelry Ai's standard Skills and your organization's custom Skills. Formerly the Prompt Library.

Custom Skill: A Skill your organization builds for its own work.

User Input: A value someone provides when running a Skill. Each Skill defines its own inputs, and a Workflow asks for the inputs of the Skills in its steps.

Document Attachment: A file attached to a Skill so every run has the same reference material.

Prompt Engineering: The practice of writing and refining instructions so you get the output you want. See Prompt Engineering.

Models​

Model provider: A company whose AI models you can use in Revelry Ai. The supported providers are OpenAI, Anthropic, Google, Groq, Mistral, SpaceXAI (Grok), and Azure (OpenAI). Your organization admin chooses which models are available.

Adaptive Thinking: The thinking mode on newer Claude models, added in v1.2.1. The model decides how much to reason for each request, within the effort level set on its model configuration. It replaced the older Extended Thinking settings. OpenAI, Google, and the other providers have their own reasoning settings.

LLM Providers page: The organization settings page where admins of enterprise organizations add their own API keys for model providers. To set this up, email support+enterprise@revelry.ai.

Knowledge and tools​

Knowledge: Your organization's searchable store of documents, and the navigation section for managing its documents, collections, and data sources. Revelry Ai searches it during chats, Skills, and Workflow runs. See Knowledge overview.

Knowledge Base: The earlier name for Knowledge.

Documents: Files stored in Knowledge. Knowledge Search can find them by meaning and by exact words.

Collections: Groupings of documents that let you scope a search to a subset of Knowledge. Collections are many-to-many: a document can belong to several collections, and a collection can hold documents from several sources. Artifacts can't be added to a collection until they're saved to Knowledge as documents.

Data Sources: External content that syncs into Knowledge as documents. The available data sources are GitHub, Notion, Google Sheets, and website crawls.

Website Crawl: A data source that crawls a website and adds its pages to Knowledge.

Chat tools: The tools chat and Workflow runs use to search Knowledge, search the web, create artifacts, and work with connected services. Several integrations below connect through MCP (Model Context Protocol) servers. See Tools in chat.

Chat Tools menu: The menu behind the + button next to the chat message input. It has four sections: Attachments (Upload New File, Select From Knowledge, and Select Artifact(s)), Built-in Tools (Knowledge, Artifacts, Web Search, and Workspace Agents), Integrations (turn connected services on or off for this chat), and Advanced. Selections last for the conversation, and the Skill Builder has the same controls. It replaced the MCP integration selector in v1.2.0.

Context button: The button below the chat box that groups a conversation's attachments and Knowledge filters once there is more than one. Click it to review them and remove one or all.

Chat Recommendations: Suggested chats, based on your recent activity, shown when you start a new chat. Turn them on or off under Chat Preferences on your Profile page. Contact your Revelry Ai team to enable.

Workspace Agents: A built-in tool that processes files in isolated Python sandboxes (E2B). You upload files for processing, and the output files are saved as downloadable artifacts. Contact your Revelry Ai team to enable.

Full-Text Search: Keyword search that matches exact terms and phrases. Artifact search uses it on artifact names and content and sorts results by relevance.

Semantic Search: Search by meaning. It finds content that matches the concepts in your question even when the wording differs. Knowledge uses semantic search alongside full-text search.

Users and permissions​

User: Anyone with access to Revelry Ai. A user can belong to several organizations and switch between them with the organization switcher at the bottom of the left navigation.

Organization Member: A user who belongs to an organization and can work in its projects. The members list shows this role as Member - Full Access.

Restricted - Group Access: A member role that sees only the projects shared with the member's groups. It's the first option when you invite new members, and the Members page shows it as Member - Group Access.

Organization Admin: A member with permission to manage the organization's settings, members, groups, and integrations. The members list shows this role as Admin.

Groups: Sets of members used for access control. Groups determine which projects, categories, documents, and data sources their members can see.

Billing and administration​

Plan: The organization's subscription: Start, Launch, Accelerate, or Scale. Start does not collect billing information. An admin upgrades to Launch, Accelerate, or Scale from Your plan, which opens from the credit balance in the left navigation.

Billing: The screen where you add a billing email and a payment method through Stripe. Upgrading from Start to Launch, Accelerate, or Scale opens it.

Terms of Service (TOS) acceptance: A step where users accept the Terms of Service before continuing: during sign-up or at their first sign-in, and when joining an organization. Contact your Revelry Ai team to enable.

API​

API Key: A credential that authenticates requests to the Revelry Ai API. Organization API keys belong to the organization. Personal API keys, added in v1.4.1, belong to one user within an organization, and anything created with a key records which key created it. See Managing API Keys.

Integrations​

GitHub: Connects in three ways. A personal or organization token adds GitHub tools to chat, and an organization-level data source syncs repositories into Knowledge. See GitHub.

Notion: A data source that syncs wikis and documentation into Knowledge. See Notion.

Google Sheets: A data source that syncs spreadsheet data into Knowledge. See Google Sheets.

Google Workspace: A personal integration for Drive, Gmail, Calendar, Docs, Sheets, and Slides in chat. Each person connects their own Google account. Chat can prepare Gmail drafts and never sends email. See Google Workspace.

Microsoft 365: An integration for Outlook Mail, Calendar, Word, Excel, and SharePoint and OneDrive through Agent 365 MCP servers. Requires Microsoft 365 Copilot licensing and Frontier Program enrollment. See Microsoft 365.

HubSpot: A CRM integration for querying contacts, companies, and deals live in chat through HubSpot's MCP server. It doesn't sync anything into Knowledge. See HubSpot.

Pipedrive: A CRM integration for reading and updating deals, contacts, organizations, activities, and notes in chat. See Pipedrive.

Procore: A construction integration for looking up projects, submittals, RFIs, subcontractors, vendors, and manpower logs in chat.

Slack: An integration for searching messages, channels, files, and members, reading and sending messages, and working with canvases in chat. An admin registers a Slack app for the organization, then each user connects their own Slack account. Contact your Revelry Ai team to enable. See Slack.

Metabase: An integration for querying dashboards, datasets, and saved questions in a partner-hosted Metabase instance (v0.60 or later) in chat. Each user signs in with their own Metabase account. Contact your Revelry Ai team to enable. See Metabase.

Amplitude: A product analytics integration for querying charts, events, and usage data in chat. Each user signs in with their own Amplitude account. Contact your Revelry Ai team to enable. See Amplitude.

Zapier: An integration that reaches thousands of apps through Zapier. Each user creates their own MCP server at mcp.zapier.com, chooses which apps and actions to expose, and connects it to Revelry Ai. Chat then runs those actions under the user's Zapier account. Contact your Revelry Ai team to enable. See Zapier.

Mapbox: An organization-level integration with six read-only location tools: points of interest, geocoding and reverse geocoding, and travel distances, times, and directions. Contact your Revelry Ai team to enable. See Mapbox.

ElevenLabs: A text-to-speech integration that creates audio narrations of artifacts, with a player built into the artifact page. Contact your Revelry Ai team to enable. See ElevenLabs.