Social media and AI agent glossary
Plain definitions of the terms you meet when you put AI agents on social media.
A
ActivityPub
ActivityPub is a W3C standard protocol for decentralized social networking that lets independent servers exchange posts, follows, likes, and other activities.
Agentic AI
Agentic AI describes AI systems that pursue a goal on their own by planning steps, calling tools, and adjusting based on results, rather than answering one prompt at a time.
AGPL (GNU Affero General Public License)
The GNU Affero General Public License (AGPL) is a copyleft open-source license that extends the GPL's source-sharing rules to software users interact with over a network.
AI agent
An AI agent is a system that uses a language model to decide on steps toward a goal and carry them out by calling tools. It acts, not only answers.
AI content generation
AI content generation is the use of machine learning models, most often large language models, to produce text, images, or other media from a prompt or source material.
AI writing assistant
An AI writing assistant is a tool that helps a person write by suggesting, drafting, editing, or rewriting text, while the person stays in control of each piece.
Alt text
Alt text is a short written description attached to an image so that people using screen readers, or anyone whose images fail to load, can understand what it shows.
API key
An API key is a secret string an application sends with each request so the API can identify the caller and decide what it is allowed to do.
API rate limit
An API rate limit is a cap on how many requests a client may send to an API within a given time window, enforced to protect the service.
Approval workflow
An approval workflow is a defined sequence of review steps that a piece of work must pass, ending with an explicit sign-off, before it is released.
AT Protocol (atproto)
The Authenticated Transfer Protocol (AT Protocol, or atproto) is an open protocol for decentralized social apps, developed by Bluesky, in which users own portable identities and signed data repositories.
Audit log
An audit log is a time-ordered record of actions taken in a system, capturing who did what, to which item, and when, so events can be reviewed later.
B
Best time to post
The best time to post is the day and hour when a specific account's audience is most likely to see and interact with its posts.
Bluesky
Bluesky is a microblogging social network built on the AT Protocol, an open system in which accounts are portable between hosting providers and users can choose their own feeds.
Brand voice
Brand voice is the consistent personality, vocabulary, and style an organization uses across its writing, so readers recognize it regardless of who wrote a given piece.
C
ChatGPT connectors
ChatGPT connectors link ChatGPT to outside apps and data sources. Custom connectors let you add your own remote MCP server so ChatGPT can call its tools.
Claude Desktop
Claude Desktop is Anthropic's desktop application for the Claude assistant on macOS and Windows, and one of the first clients to support connecting to MCP servers.
Content approval
Content approval is the step in which a designated person reviews a piece of content and signs off on it before it is published.
Content calendar
A content calendar is a schedule that lists what will be published, on which channels, and when, so a team can plan ahead and avoid gaps or collisions.
Content repurposing
Content repurposing is turning one existing piece of content into new formats or versions for other channels and audiences.
Cross-posting
Cross-posting is publishing the same piece of content, sometimes adapted, to more than one social network or account.
Cursor (IDE)
Cursor is an AI code editor, built as a fork of Visual Studio Code, that adds chat, code generation, and an agent mode that can edit files, run commands, and call MCP tools.
D
E
Engagement rate
Engagement rate is the share of people who interacted with a post, calculated as total engagements divided by impressions or by followers, expressed as a percentage.
Evergreen content
Evergreen content is material that stays accurate and useful long after it is published, so it can be shared again without going stale.
F
G
H
Hash chain
A hash chain is a sequence of records in which each record includes a cryptographic hash of the previous one, so altering any earlier record breaks every link after it.
Human in the loop (HITL)
Human in the loop (HITL) is a design where a person reviews, corrects, or approves an automated system's output before it takes effect.
I
K
L
Large language model (LLM)
A large language model (LLM) is a neural network trained on large amounts of text to predict the next token, which lets it generate and transform language.
Link shortener
A link shortener is a service that replaces a long URL with a short one that redirects to the original, often while counting clicks.
LinkedIn Company Page
A LinkedIn Company Page, now officially called a LinkedIn Page, is an organization's profile on LinkedIn, managed by one or more admins, where it posts updates and lists jobs.
M
Mastodon
Mastodon is free, open-source social network software run on thousands of independent servers that connect to each other through the ActivityPub protocol.
MCP server
An MCP server is a program that exposes tools, resources, and prompts to AI applications through the Model Context Protocol. The AI client connects to it and calls those capabilities on the user's behalf.
Model Context Protocol (MCP)
The Model Context Protocol (MCP) is an open standard for connecting AI applications to external tools and data. It defines how a client discovers and calls capabilities that a server exposes.
N
O
OAuth
OAuth is an open authorization standard that lets one application act on your account at another service, with limited permissions, without ever receiving your password.
Open source
Open source software is software whose source code is published under a license that lets anyone use, study, modify, and redistribute it.
P
Posting frequency
Posting frequency is how often an account publishes on a given network, usually measured per day or per week.
Principle of least privilege
The principle of least privilege says every user, program, or agent should have only the access it needs to do its task, and nothing more.
Prompt injection
Prompt injection is an attack in which text supplied to a language model, directly or hidden in data it reads, overrides the instructions the developer or user intended.
R
Reach
Reach is the number of unique accounts or people who saw a post at least once during a given period.
Retrieval-augmented generation (RAG)
Retrieval-augmented generation (RAG) is a technique in which a system first retrieves documents relevant to a request and then gives them to a language model as context for its answer.
S
Self-hosting
Self-hosting means running a piece of software on infrastructure you control, such as your own server or cloud account, rather than using the vendor's hosted service.
Social listening
Social listening is the practice of tracking public online conversations about a brand, product, competitor, or topic, and analyzing them for trends, sentiment, and issues.
Social media automation
Social media automation is the use of software to carry out repetitive social media tasks, such as scheduling, cross-posting, and reporting, without a person doing each step.
Social media manager
A social media manager is the person responsible for an organization's social accounts: planning content, writing or commissioning posts, publishing, replying, and measuring results.
Social media scheduler
A social media scheduler is software that stores posts, queues them for specific times, and publishes them to connected social accounts through each network's API.
stdio (MCP transport)
stdio (standard input/output) is the Model Context Protocol's local transport: the client launches the server as a child process and exchanges newline-delimited JSON-RPC messages over its stdin and stdout.
Streamable HTTP (MCP transport)
Streamable HTTP is the Model Context Protocol's standard transport for remote servers: the client sends JSON-RPC messages as HTTP POST requests to a single endpoint, and the server replies with JSON or a stream of server-sent events.
T
Thread
A thread is a series of connected posts, each replying to the one before, that together read as a single longer piece.
Tool calling
Tool calling, also called function calling, is the way a language model asks the surrounding application to run a named function with structured arguments and then uses the result.
Trust ladder
A trust ladder is a staged permission model in which an AI agent starts with minimal autonomy and is promoted to more only after a record of good work, and demoted when it errs.
U
W
X
Z
Let your agents post. Keep the final say.
Connect your first agent in minutes. Every draft waits in your inbox until you approve it.
14 days free. No card required.