Beav Docs
Beav documentation
Start with the desktop app and browser extension, then learn web access, capture, knowledge, content creation, media processing, and external Agent connections.
Beav is a local-first AI workspace for creator content and assets. It keeps web pages, comments, images, videos, local files, and finished drafts in one workspace, so every source stays searchable, citable, and reusable.
Start here
Using Beav for the first time? Finish these steps in order:
- Install the Beav desktop app.
- Install the Chrome / Edge browser extension.
- Save your first web page.
- Turn captured sources into topics and drafts.
Common entry points
| Goal | Guide |
|---|---|
| Learn about spaces, navigation, account profiles, and interface language | Get to know the workspace |
| Reference material in chat, use voice input, and use the right workspace | Chat and the right workspace |
| Read the daily brief, confirm today's plan, and review AI output | Daily brief and today's plan |
| Edit manuscripts, save publishing packages, and pick them up on your phone by scanning | Manuscripts, packages, and publishing |
| Have the AI remember your preferences and writing style | Memory and creative style |
| Save web pages, Xiaohongshu, YouTube, images, or selected text | Capture pages and platform content |
| Save by pasting a link, subscribe to benchmark creators, and sync your own account | Links, subscriptions, and account profiles |
| Break down videos and notes, and go from material to finished draft | From capture to draft |
| Organize folders, selections, and source relationships | Organize knowledge, assets, and manuscripts |
| Use image and video templates | Creation templates |
| Cut talking-head videos, add subtitles and animation, and export video | Video editing and smart talking-head cuts |
| Extract video scripts, transcribe, dub, and make Live Photos | Built-in mini tools |
| Browse trends and discuss content angles | Explore trends |
| Connect an Eagle library and reference images | Import and use an Eagle library |
| Use Beav from a browser, a phone, or a server | Web access |
| Let Beav drive real, already signed-in browser pages | Browser control |
| Call third-party APIs or automate a repeatable workflow | Environment variables and scripts |
| Install Skills or plugins, or connect MCP services | Skills, plugins, and MCP |
| Let Codex or WorkBuddy use Beav | Connect an external Agent |
| Connect third-party models or use a redemption code | Account, credits, and model settings |
| Fix install, connection, or long-task problems | Troubleshooting |
How Beav organizes creation
The browser extension, phone uploads, pasted links, and local imports bring material into your account's space. The knowledge base keeps reference content and originals, the asset library keeps reusable images, videos, copy, and music, and the account profile keeps your own account's notes and data. In chat, the AI works from this real material to shape topics and write drafts, producing manuscripts you can keep editing and publishing packages for each platform; generated images, videos, and audio return to the same space.
An external Agent that connects to Beav should follow the same path: read the real sources and the current task first, then create a draft or media task that follows the request, and finally save the results back into Beav — instead of leaving behind a block of chat text nobody can reuse.
Data and permissions
By default Beav keeps workspace data on your machine. Browser capture, media generation, and third-party model calls reach the relevant page or service only for actions you start. Paid generation, browser control, deletion, publishing, and external exports from an external Agent require approval; see Capabilities and approvals for the full rules.
Mini Apps
Mini Apps are local apps installed inside Beav. They keep their own interface state and, once the user grants permission, can call Beav's capture, knowledge base, asset library, media processing, AI, and automation capabilities.
Read the Mini App development guide for the package structure, the window.redbox SDK, every capability name, the security boundary, and the release checklist.