Skip to main content
Use this glossary to understand the terminology used across Runable. Definitions explain what each term means in the context of the Runable platform. Select a linked term to open its detailed guide. A · B · C · D · E · F · I · L · M · O · P · R · S · T · V · W

​
A

Agent — The AI system that interprets your request, chooses an approach, uses available tools, and returns a completed result. In Runable, the agent can answer questions, work with files, write and execute code, generate media, and build deployable projects. Agent Mode — Runable’s default mode for creating and executing work. Agent Mode gives the agent access to a private sandbox and tools for code execution, file operations, web research, media generation, and multi-step tasks. AI Audio — Runable’s audio creation toolkit for text-to-speech, multi-speaker dialogue, music, sound effects, voice cloning, voice swapping, dubbing, and transcription. AI Canvas — A freeform visual workspace for generating, uploading, arranging, editing, and exporting images and videos. Media created with AI Images, AI Videos, and AI Carousel can appear on the canvas. AI Carousel — A Runable feature for creating swipeable, multi-page visual posts from a prompt or template. Carousels can be designed for platforms such as Instagram, LinkedIn, and X. AI Chat — The central conversational interface in Runable. AI Chat supports Agent Mode for building, Chat Mode for conversation, and Plan Mode for planning before execution. AI Images — Runable tools for generating images from text and editing existing images. Supported workflows include background removal or replacement, upscaling, text editing, and targeted region editing. AI Reports — A Runable feature for producing structured, multi-page reports and analyses from a prompt. Reports can include researched content, headings, tables, formatted sections, version history, and PDF export. AI Slides — A Runable feature for generating presentation decks from a prompt. The agent can create the content, structure, layout, and visual design, and the result can be exported to presentation formats. AI Spreadsheets — A Runable feature for generating, viewing, editing, and exporting Excel-compatible workbooks. Spreadsheets can include formulas, charts, conditional formatting, data validation, and multiple sheets. AI Videos — Runable tools for generating videos from text or images and editing existing clips. Workflows include natural-language editing, background removal, motion control, and combining clips into a timeline. AI Websites — A Runable feature for building and deploying full-stack websites and web applications through conversation. A project can include a database, file storage, payments, authentication, a custom domain, and analytics. Artifact — A file or interactive result created during a Runable task. Examples include a website, presentation, report, spreadsheet, image, video, audio file, code project, or downloadable document. Authentication — The process of verifying a user’s identity and maintaining their signed-in session. A Runable-built website can include sign-up, sign-in, magic-link, or supported social login flows.

​
B

Branch — A separate copy of a conversation and its working state created from an earlier checkpoint. Branching lets you try a different direction without overwriting the original version.

​
C

Chat Mode — A conversation-focused Runable mode for questions, analysis, brainstorming, and file review. Unlike Agent Mode, Chat Mode does not create a sandbox or execute code to build artifacts. Checkpoint — A saved state in an Agent Mode conversation. Checkpoints make it possible to restore earlier work with rollback or create an alternative version with branching. Context Window — The amount of information an AI model can consider at one time, including instructions, conversation history, and attached content. Larger context windows can help with long documents and complex projects. Credit — A Runable usage unit consumed when models and tools perform work. Credit use varies with the selected model tier, the type of generation, and the complexity of the task. Custom Domain — A domain or subdomain owned by you and connected to a Runable-deployed website, such as www.example.com. Runable handles the connection and SSL setup for supported deployments. Custom Instructions — Persistent directives that tell the Runable agent how to respond or work across future conversations. They can specify tone, formatting rules, business context, preferred technologies, or other standing requirements. Custom Skill — An agent skill created or uploaded by a user for a specific workflow. Custom skills can encode repeatable instructions, tools, templates, and process rules for your own work.

​
D

Database — Structured storage used by an application for records such as users, products, posts, orders, or settings. Runable websites can use an edge-hosted SQLite database for application data. Deployment — The process of publishing a website or application to a live environment where other people can access it. A deployment turns the project in the agent’s workspace into a public, running product. Dubbing — The process of replacing spoken content in an audio or video file with speech in another language. AI dubbing can automate transcription, translation, and regenerated speech.

​
E

Environment Variable — A configuration value supplied to an application outside its source code. Environment variables are commonly used for API keys, secrets, service URLs, and settings that differ between environments. Export — A downloadable copy of a Runable artifact in a supported file format. Examples include PDF reports, PPTX presentations, XLSX or CSV spreadsheets, ZIP archives, images, videos, and audio files.

​
F

File Storage — Storage for application files such as images, documents, user uploads, and generated assets. File storage is designed for objects that do not belong in ordinary database rows. Full-Stack Web App — A web application that includes both the user-facing interface and the server-side systems behind it. A full-stack app may include business logic, authentication, a database, file storage, payments, and deployment infrastructure.

​
I

Image-to-Video — A video-generation workflow that turns a still image into an animated clip. The prompt describes the motion, camera behavior, atmosphere, or changes that should occur. In-App Purchase — A purchase made inside a mobile app through Apple’s App Store or Google Play. Common types include subscriptions, one-time feature unlocks, and consumable credits or items.

​
L

Large Language Model (LLM) — An AI model trained to understand and generate language. In Runable, an LLM can interpret prompts, reason about tasks, write content or code, and help decide which tools to use.

​
M

Mark & Edit — A precision image-editing workflow in which you mark one or more regions and describe the change required in each region. Runable applies the requested edits while preserving the rest of the image as much as possible. Memory — Account-level information the Runable agent remembers across conversations. Memory can store reusable preferences and context such as brand details, tone of voice, business information, and technical defaults. Mobile App — An application designed to run on a mobile device, typically distributed through the Apple App Store or Google Play. Runable can help create, preview, test, update, and prepare mobile apps for store deployment. Model Tier — A Runable selection that balances speed, reasoning quality, and credit use. Agent Mode currently organizes models into Lite for faster simple tasks, Pro for general work, and Max for deeper reasoning and complex projects. Multimodal AI — AI that can work with more than one type of information, such as text, images, documents, audio, video, or code. Multimodal workflows allow a prompt and attached files to be considered together. Music Generation — The creation of original music from a text description. A prompt can specify elements such as genre, mood, instruments, tempo, and intended use.

​
O

Object Storage — A storage system optimized for files and other unstructured objects rather than rows and columns. Runable websites can use object storage for uploads, images, documents, and media assets. Official Skill — A skill from the Runable skill library that is built and maintained by the Runable team or a verified contributor. Official skills can be enabled or disabled but are not edited like custom skills. Over-the-Air Update (OTA Update) — A software update delivered to an installed mobile app without requiring the user to download a completely new store release. OTA updates are commonly used for compatible fixes and content or interface changes.

​
P

Plan Mode — A Runable mode that adds a planning stage before execution. The agent researches the request, gathers missing requirements, and prepares a plan for review before it starts building. Prompt — The instruction or request you give an AI system. A useful prompt states the desired outcome, relevant context, constraints, source material, and the format of the expected result. Push Notification — A message sent by a mobile app to a user’s device outside the active app session. Push notifications can deliver reminders, updates, alerts, or links that open a specific screen.

​
R

Rollback — Restoring an Agent Mode conversation and its working files to an earlier checkpoint. Rollback is useful when a later change should be discarded rather than preserved as a separate version. Runable — An AI work platform for building, running, and growing projects through conversation. Runable combines AI chat, code execution, content creation, media generation, websites, apps, skills, and connected workflows. RunClaw — A way to use the Runable agent from supported messaging platforms. It lets you send requests and receive results without starting every interaction in the Runable web application.

​
S

Sandbox — A private, isolated environment where the Runable agent can write files, execute code, install packages, and run commands without operating directly on your computer. Each Agent Mode conversation has its own working environment. Search Engine Optimization (SEO) — The practice of improving a website so search engines can understand, index, and rank its pages for relevant searches. Common SEO elements include page titles, descriptions, canonical URLs, internal links, structured content, and accessible public pages. Skill — A reusable workflow that extends how the Runable agent performs a particular type of task. A skill can contain instructions, tools, templates, and rules that the agent loads when the user’s request matches it. Snapshot — A saved version of content or project state at a particular moment. Snapshots support version history by making it possible to inspect or restore an earlier version. Sound Effect Generation — The creation of an audio effect from a written description. Prompts can describe the sound source, environment, intensity, duration, and whether the result should loop. Speech-to-Text (STT) — Technology that converts spoken audio into written text. STT is the underlying capability used for transcription, captions, searchable speech, and speaker-aware analysis. SSL Certificate — A digital certificate that enables an encrypted HTTPS connection between a website and its visitors. SSL protects data in transit and helps browsers verify the site’s domain identity. Stripe Integration — A connection between a Runable-built website and Stripe for accepting supported payments. The integration can be used for payment flows such as one-time charges or subscriptions.

​
T

Text-to-Image — An image-generation workflow that creates a new visual from a written prompt. The prompt can describe the subject, composition, style, lighting, colors, aspect ratio, and other visual requirements. Text-to-Speech (TTS) — Technology that converts written text into synthesized spoken audio. TTS systems can offer different voices, languages, speaking styles, and delivery settings. Text-to-Video — A video-generation workflow that creates a moving clip from a written description. The prompt can define the scene, subject, movement, camera direction, mood, and visual style. Transcription — The process and resulting text produced by converting speech in an audio or video file into written words. A transcription may also include speaker labels and detected audio events.

​
V

Voice Cloning — The creation of a synthetic voice based on audio samples of a speaker. A cloned voice can then be used to generate new speech while approximating the source voice’s characteristics. Voice Swap — Replacing the voice in an existing recording while preserving other aspects of the performance, such as pacing, timing, and emotional delivery.

​
W

Website Analytics — Data that helps you understand how people use a deployed website. Common measurements include page views, visitors, traffic behavior, and tracked events.

​
Continue learning