> AI agents: this is one page from Mammoth Analytics documentation. The index of all pages as Markdown is https://mammoth.io/docs/llms.txt. Append `.md` to any docs URL, or send `Accept: text/markdown`, to get Markdown. # Mammoth glossary: what each Mammoth term means > Plain definitions of Mammoth terms: Workspace, Project, Dataset, Batch, View, Pipeline, Task, Automation, AI Agent, Ask mode, Live link, Connector and more. Mammoth's terms are defined below in alphabetical order, from Activity Log to Your agent. Each entry says what the term is and, where it matters, where you find it in the app and who can use it. It links to the page that covers it in full. To see how the main terms fit together, read [Core ideas](https://mammoth.io/docs/learn/start/core-ideas/). ## A ### Activity Log The Activity Log is Mammoth's audit trail. It records the actions people take in your workspace, from creating datasets and editing pipelines to changing Automations and managing users. Workspace Owners and Workspace Members open it from the **Activity Log** tab in **Monitor**. How far back it goes depends on your plan. See [Activity log](https://mammoth.io/docs/learn/automate/activity-log/). ### AI Agent The AI Agent is a chat inside the Mammoth web app. You say what you want in plain words, and it answers questions, makes charts, cleans and combines data, builds dashboards and connects sources, with your own access. Open it from **Agent** in the left navigation. A Workspace Owner turns it on or off for the whole workspace. The AI Agent, including Ask mode and Today, is available on Enterprise plans. See [AI Agent](https://mammoth.io/docs/learn/agent/). ### API token An API token lets a script, an integration or a CI job call Mammoth. It starts with `mm_` and works in every project of its workspace unless you scope it to one project. Workspace Owners and Workspace Members create tokens on the **API Keys** page of Workspace Settings, and the secret is shown once. AI assistants and the CLI sign in instead and need no token. See [API keys](https://mammoth.io/docs/learn/workspace/api-keys/) and [Authentication](https://mammoth.io/docs/build/auth/). ### Ask mode Ask mode is the read-only side of the app, built around the AI Agent. It answers, makes charts and writes summaries from the data you can read, and never changes data. Its navigation includes **Today**, **Chats**, **Explore**, **Snippets** and **Your agent**, plus **Dashboards** when boards are shared with you. When a Workspace Owner turns Ask mode on, people who can edit a project get an **Ask** and **Build** switch at the top of the left navigation. Available on Enterprise plans. See [Ask mode](https://mammoth.io/docs/learn/agent/ask-mode/). ### Auto-sync Auto-sync is a toggle that controls whether a View or an export updates automatically when its upstream data changes. Mammoth has two independent layers: **View Auto-sync** re-runs a View's pipeline when its source data changes, and **Export Auto-sync** sends a View's new output to its destination after each run. When Auto-sync is off, changes stop at that point, and the item is listed in Monitor's **Paused** tab, where you can turn Auto-sync back on. See [View auto-sync](https://mammoth.io/docs/learn/organise/view-auto-sync/), [Send to destinations](https://mammoth.io/docs/learn/share-and-export/send-to-destinations/), and [Live links](https://mammoth.io/docs/learn/share-and-export/live-links/). ### Automation An Automation runs recurring work in a project without manual steps. Mammoth has five Automation types: Dataset Refresh, File Collection, Data Consolidation, Scheduled Export and Messaging. You create and manage them from the **Automations** tab in the Data Library. See [What you can automate](https://mammoth.io/docs/learn/automate/) and [Manage automations](https://mammoth.io/docs/learn/automate/manage/). ## B ### Batch A Batch is one load of data inside a Dataset. Each time data is added to a Dataset, it creates a Batch with its own timestamp and row count. A Batch can be suspended (left out without deleting it) or deleted, and new data can be added to the existing Batches or replace them. See [Core ideas](https://mammoth.io/docs/learn/start/core-ideas/) and [Batches](https://mammoth.io/docs/learn/get-data-in/batches/). ### Build Build is the part of the app where you change data: the Data Library, pipelines, dashboards and Automations. When Ask mode is on, the **Ask** and **Build** switch at the top of the left navigation moves between the two. The AI Agent in Build can change data in projects you can edit. See [Ask mode](https://mammoth.io/docs/learn/agent/ask-mode/). ## C ### Checkpoint A Checkpoint is a pipeline control that pauses the pipeline at a specific step to notify your team or wait for approval before continuing. Use it when a person needs to review the data before the pipeline proceeds. Checkpoints waiting for approval appear in Monitor's **Needs Attention** tab. See [Pause a pipeline with a Checkpoint](https://mammoth.io/docs/learn/clean-and-shape/checkpoints/). ### CLI The Mammoth CLI (`mammoth`) runs Mammoth from a terminal or a script, one command at a time, and coding agents such as Claude Code can run it for you. It works on every plan. Sign in with `mammoth auth login`. See [CLI](https://mammoth.io/docs/build/cli/) and [Mammoth for coding agents](https://mammoth.io/docs/agents/). ### Collection A collection is a set of dashboards you share as one. Everyone it is shared with sees every dashboard in it, including ones added later. A collection is a sharing unit, not a folder: a dashboard can sit in several collections. Collections belong to one project, and its Project Admins and Project Editors manage them. See [Collections](https://mammoth.io/docs/learn/share-and-export/collections/). ### Connection A connection is a saved sign-in to a source or destination: a database login, an app you signed in to, or an API key. It belongs to a project, and the **Connections** tab beside **Datasets** lists every connection in the project with what uses it. Project Admins create, edit and delete connections. Project Editors can start a dataset from one. See [Manage saved connections](https://mammoth.io/docs/learn/get-data-in/connections/). ### Connector A Connector is the type of source or destination Mammoth can talk to, such as PostgreSQL, Snowflake, Salesforce, Google Drive or a REST API. You pick one in the **Connectors** launcher, and connecting it creates a connection. See [Connect a database or app](https://mammoth.io/docs/learn/get-data-in/connect-a-source/). ### Connector workspace The connector workspace is where you add data from a connector: you connect the source, pick the tables and columns you want, and save them as a dataset. A chat beside it can build the connection for you. See [Connect a database or app](https://mammoth.io/docs/learn/get-data-in/connect-a-source/). ### Context A context is a reusable bundle of knowledge, such as goals, definitions and reference files, that Mammoth grounds on when it writes a dashboard's text and answers its questions. You keep contexts in the **Context library** on the **Context** tab of **Dashboards** and attach them to any dashboard. See [Give the AI context](https://mammoth.io/docs/learn/dashboards/context/). ### Custom role A custom role is a workspace or project role you build from Mammoth's permission list, with **New Role** in **Roles & Permissions** in Workspace Settings. Custom roles are included on the Pro plan. See [Members, roles and custom roles](https://mammoth.io/docs/learn/workspace/members-and-roles/). ## D ### Dashboard A Dashboard is an interactive set of charts, numbers and text that the AI builds from data you pick in your project, after you describe what you want. Each item in the **Dashboards** library is a Dashboard, Doc, Slides deck, Q&A page or Explore page. You can publish it to a link and share it. See [Dashboards](https://mammoth.io/docs/learn/dashboards/), [Create a dashboard](https://mammoth.io/docs/learn/dashboards/create/) and [Publish and share a dashboard](https://mammoth.io/docs/learn/share-and-export/share-a-dashboard/). ### Dashboard Editor and Dashboard Viewer Dashboard Editor and Dashboard Viewer are the two roles on a single dashboard. A Dashboard Editor can modify the dashboard, and a Dashboard Viewer can only view it. See [Members, roles and custom roles](https://mammoth.io/docs/learn/workspace/members-and-roles/). ### Data Check A Data Check is an automated quality gate attached to a pipeline step. After the step runs, the Data Check evaluates the output against rules you define, such as row count, null percentage, unique values, value range or a custom prompt, and either pauses the pipeline or flags the problem and continues. See [Remove duplicates and check data](https://mammoth.io/docs/learn/clean-and-shape/duplicates-and-checks/). ### Data Consolidation Data Consolidation is an Automation type that takes files added to a source folder and puts their data into a destination Dataset. It runs whenever a new file lands in the folder, with no schedule. It either adds the rows to the existing data (**Add to existing data**) or replaces the data with the new file's. See [Combine incoming data](https://mammoth.io/docs/learn/automate/consolidate/). ### Data Library The Data Library holds a project's Datasets and Folders, alongside its connections, its **Automations** tab and the reusable pipeline components called Parameters and Snippets. See [Find your way around the Data Library](https://mammoth.io/docs/learn/organise/data-library/). ### Data Quality Report The Data Quality Report is an AI assessment of a View's data health against six DAMA quality dimensions. It gives an overall quality score, lists critical issues and warnings, and recommends pipeline tasks to fix them. Generate it with **View Data Quality** in the footer of a View. See [Data quality report](https://mammoth.io/docs/learn/explore/data-quality/). ### Dataset A Dataset is the result of bringing data into Mammoth. It holds your working data with a schema, metadata and a collection of Batches. See [Core ideas](https://mammoth.io/docs/learn/start/core-ideas/) and [The dataset panel](https://mammoth.io/docs/learn/organise/dataset-panel/). ### Dataset Refresh Dataset Refresh is an Automation type that fetches the latest data for a connected Dataset on a schedule or when you run it. It works on Datasets connected to an API or database, and on Datasets that follow the link they came from. Uploaded files and files from drives can't be refreshed. See [Refresh a dataset](https://mammoth.io/docs/learn/automate/refresh-a-dataset/). ### Draft mode Draft mode is a pipeline setting that queues your task changes as pending updates instead of running them immediately. The queued changes run together when you click **Apply now**. Turn it on from the **More Pipeline options (...)** menu in the Pipeline panel. See [How the pipeline works](https://mammoth.io/docs/learn/clean-and-shape/how-the-pipeline-works/#apply-several-tasks-at-once-with-draft-mode). ## E ### Email group An email group is a named list of email addresses. Type its name in any invite field to invite everyone in it. You manage groups in **Email groups** in Workspace Settings. See [Workspace settings](https://mammoth.io/docs/learn/workspace/settings/). ### Email updates Email updates send the people a dashboard or collection is shared with its headline numbers, when new data is published or on a schedule. You set them up in the **Email updates** section of the share settings. They go to signed-in viewers only. See [Email updates](https://mammoth.io/docs/learn/share-and-export/email-updates/). ### Embed Embedding puts a published dashboard, one of its pages or one chart in a page you control. In the **Embed** tab of the Share panel you choose who sees it: anyone (over the public link), **Basic** (a key in the page) or **Advanced** (a token your server signs for each viewer). Embedding is available on every plan. See [Publish and share a dashboard](https://mammoth.io/docs/learn/share-and-export/share-a-dashboard/) and [Embedding](https://mammoth.io/docs/build/embedding/). ### Explore Cards Explore Cards are interactive cards that show what is inside a column: the values of a text column, the distribution of a numeric column, or the timeline of a date column. Clicking values on a card filters the grid. See [Explore cards](https://mammoth.io/docs/learn/explore/explore-cards/). ### External API key An external API key is your own key for an AI provider, such as OpenAI or Anthropic, that Generative AI steps can use. Workspace Owners add them in **External API keys** on the **API Keys** page of Workspace Settings. Without one, Generative AI uses a key Mammoth provides. See [API keys](https://mammoth.io/docs/learn/workspace/api-keys/) and [Use AI on your columns](https://mammoth.io/docs/learn/clean-and-shape/ai-on-columns/). ## F ### File Collection File Collection is an Automation type that collects files from folders in cloud platforms such as Dropbox, Google Drive, SFTP or SharePoint into a destination folder in Mammoth. It runs on a schedule or with **Run now**, and it is a plan feature. See [Collect files](https://mammoth.io/docs/learn/automate/collect-files/). ### Folder A Folder organizes Datasets in the Data Library. You can move Datasets and Folders into other Folders, and a Data Consolidation Automation can watch a Folder for new files. See [Folders](https://mammoth.io/docs/learn/organise/folders/). ## G ### Guest A guest is someone a Workspace Owner invites to ask the AI Agent questions, with **Invite to a room** on the **AI Agent** page of Workspace Settings. Guests join free and never take a seat. They see only Ask, read only the projects and files picked for them, and have a monthly question limit. Available on Enterprise plans. See [AI Agent access](https://mammoth.io/docs/learn/workspace/ai-agent-access/). ## I ### Insights Insights, also called View Insights, are short business observations that Mammoth's AI writes from the statistical profile of a View. Click **View Insights** in the action bar to generate them. See [View insights](https://mammoth.io/docs/learn/explore/view-insights/). ## L ### Live link A Live Link is a URL tied to a View that always serves the View's latest output, as CSV. Create one from **Export** in the View toolbar. It sits as a task at the end of the pipeline, and each View can have one. See [Live links](https://mammoth.io/docs/learn/share-and-export/live-links/). ## M ### MCP server The Mammoth MCP server, at `https://mcp.mammoth.io/mcp`, lets your own AI client, such as Claude or ChatGPT, explore, clean, transform and export your Mammoth data. You sign in with your Mammoth account, and every tool call runs with your permissions. It works on every plan and is separate from the in-app AI Agent. See [Mammoth MCP server](https://mammoth.io/docs/agents/mcp/). ### Messaging Messaging is an Automation type that emails people on a schedule, with Dataset Views attached as CSV files. It has no manual-only option, and your plan sets how many Messaging automations you can have. See [Send messages](https://mammoth.io/docs/learn/automate/messages/). ### Metrics Metrics are single-value calculations such as total revenue, average deal size or defect rate, built from a column in the Metrics panel. They update as you filter your data in Explore Cards. See [Metrics](https://mammoth.io/docs/learn/explore/metrics/). ### Monitor Monitor shows what is happening in a project: items that need your action, pipelines and Automations that are running, items with Auto-sync paused, and the Activity Log. Open it from **Monitor** in the left sidebar. See [Monitor](https://mammoth.io/docs/learn/automate/monitor/). ## N ### Notifications The notifications bell at the bottom of the left sidebar tells you when something you started or own finishes or fails, such as a download, an export, a data pull or a data check. See [Notifications](https://mammoth.io/docs/learn/automate/notifications/). ## O ### OAuth app An OAuth app is an app registered in your workspace so people can connect their Mammoth account to it with Sign in with Mammoth. Workspace Owners register and delete OAuth apps in **OAuth apps** on the **API Keys** page of Workspace Settings. See [API keys](https://mammoth.io/docs/learn/workspace/api-keys/). ## P ### Parameter A Parameter is a named value, either a number, a text string or a date, that you reference in pipeline conditions, filters and calculations. Change the value once and every pipeline that uses it updates. See [Parameters and snippets](https://mammoth.io/docs/learn/organise/parameters-and-snippets/). ### Pipeline A Pipeline is the ordered list of Tasks that transform the data in a View. It runs top to bottom, re-runs when the data refreshes, and every step can be reviewed, edited, reordered or removed. See [How the pipeline works](https://mammoth.io/docs/learn/clean-and-shape/how-the-pipeline-works/). ### Project A Project is an organizational container inside a Workspace. It groups related Datasets, Automations and dashboards, and controls which people can access them. See [Projects](https://mammoth.io/docs/learn/workspace/projects/). ### Project Admin, Project Editor and Project Viewer Project Admin, Project Editor and Project Viewer are the built-in project roles. A Project Admin manages the project's settings, members and connections, as well as everything an Editor can do. A Project Editor works with the project's data and pipelines. A Project Viewer can only view the project's Data Library items. See [Projects](https://mammoth.io/docs/learn/workspace/projects/) and [Members, roles and custom roles](https://mammoth.io/docs/learn/workspace/members-and-roles/). ## S ### Scheduled Export Scheduled Export is an Automation type that runs export steps you already added to Views at fixed times, whether or not the data changed. Each export runs to its own destination. See [Export on a schedule](https://mammoth.io/docs/learn/automate/scheduled-export/). ### Skills Skills are the sets of steps and tools the AI Agent uses for one kind of job, such as reading your data, connecting a source or drawing a chart. They are built into Mammoth: you don't pick or install them, and the agent loads the one it needs. See [What the AI Agent can do](https://mammoth.io/docs/learn/agent/skills/). ### Snippet A Snippet is a reusable block of SQL code that you embed in tasks using double-brace syntax. Where a Parameter holds a single value, a Snippet holds logic. See [Parameters and snippets](https://mammoth.io/docs/learn/organise/parameters-and-snippets/). The **Snippets** entry in Ask's navigation is a different thing: there, a snippet is a canvas of charts, numbers and notes saved from your chats. See [Results and the data canvas](https://mammoth.io/docs/learn/agent/results-and-canvas/). ## T ### Task A Task is one transformation step in a Pipeline, such as a filter, a join or a calculation. Each Task does one operation, is set up in its own panel, and can be previewed. See [Core ideas](https://mammoth.io/docs/learn/start/core-ideas/). ### Today Today is the home page of Ask mode. Your agent writes it from the data Ask answers from: the numbers that matter, what changed and anything far from usual. It is written again when your data refreshes. Available on Enterprise plans. See [Today](https://mammoth.io/docs/learn/agent/today/). ### Trash Trash is where deleted Datasets, Views, Automations, Dashboards and Folders wait before being permanently removed. Items stay recoverable for 30 days by default, and each project has its own Trash. See [Get things back from Trash](https://mammoth.io/docs/learn/organise/trash/). ## V ### View A View is where you work with your data. It shows the data in an interactive grid, with Explore Cards, a Metrics panel and a Pipeline for building transformations. One Dataset can have many Views, and Views do not change the source data. See [Core ideas](https://mammoth.io/docs/learn/start/core-ideas/) and [Views: create, rename and delete](https://mammoth.io/docs/learn/organise/views/). ## W ### Workspace A Workspace is your organization's whole Mammoth environment. It is where the plan and billing are managed, users are invited and given roles, and all Projects live. See [Core ideas](https://mammoth.io/docs/learn/start/core-ideas/) and [Manage your workspace](https://mammoth.io/docs/learn/workspace/). ### Workspace Owner and Workspace Member Workspace Owner and Workspace Member are the built-in workspace roles. A Workspace Owner manages users, roles, billing and workspace settings. A Workspace Member can create projects, create API tokens and view the Activity Log, and reaches a project only after being added to it. See [Members, roles and custom roles](https://mammoth.io/docs/learn/workspace/members-and-roles/). ## Y ### Your agent **Your agent** is a page in Ask mode that shows what the AI Agent knows about you, what it answers from and what it sends you, including your alerts. Each person has their own. Available on Enterprise plans. See [Your agent](https://mammoth.io/docs/learn/agent/your-agent/). ## Related - [Core ideas](https://mammoth.io/docs/learn/start/core-ideas/): how workspaces, projects, datasets, Views and pipelines fit together. - [FAQ](https://mammoth.io/docs/learn/help/faq/): short answers to common "how do I" questions. - [Limits and quotas](https://mammoth.io/docs/learn/help/limits/): the numbers behind the terms, such as Trash time and plan allowances. - [Troubleshooting](https://mammoth.io/docs/learn/help/): find the right guide when something goes wrong. --- Source: https://mammoth.io/docs/learn/help/glossary.md ยท Updated: 2026-11-08