Bee Box documentation · directory: https://beebox.run/docs/ · index: https://beebox.run/llms.txt · root: https://beebox.run/llms.txt # What it can do A **card** is one file with a structured header that gets checked against its type; a **box** is the directory of them. One page per capability is in [capabilities/index.md](https://beebox.run/docs/capabilities/index.md). ## Capability areas - [Gmail](https://beebox.run/docs/capabilities/gmail.md): sync watched threads in, draft replies you send yourself. - [Calendar](https://beebox.run/docs/capabilities/calendar.md): two-way Google Calendar sync. - [Drive](https://beebox.run/docs/capabilities/drive.md): mirror Google Docs, Sheets, and folders both ways. - [Telegram](https://beebox.run/docs/capabilities/telegram.md): chat with the box from a messaging app. - [The web interface](https://beebox.run/docs/capabilities/web-interface.md): dashboard, browser, questions, map, history, settings, and a display for each kind of card, on desktop and phone. - [Chat](https://beebox.run/docs/capabilities/chat.md): the central web conversation, with resumable threads. - [Voice](https://beebox.run/docs/capabilities/voice.md): dictation, transcription, and spoken replies. - [Phone capture](https://beebox.run/docs/capabilities/phone-capture.md): record, photograph, and upload from a phone. - [Triage](https://beebox.run/docs/capabilities/triage.md): classify and file what arrives, with confidence levels. - [Questions](https://beebox.run/docs/capabilities/questions.md): the agent asks instead of guessing. - [Schedules](https://beebox.run/docs/capabilities/schedules.md): recurring runs and agent-set timers. - [Procedures](https://beebox.run/docs/capabilities/procedures.md): declarative multi-step workflows. - [Dump it in now, shape it later](https://beebox.run/docs/capabilities/shape-it-later.md): put things in before deciding their structure; what does not fit stays in your words and can be reshaped later. - [It keeps itself coherent](https://beebox.run/docs/capabilities/integrity.md): links parsed and checked, references rewritten on a move, cards validated at several layers, so the box stays navigable as it grows. - [Provenance](https://beebox.run/docs/capabilities/provenance.md): your exact words kept as quotes, and facts that point back at their source. - [Views](https://beebox.run/docs/capabilities/views.md): pages and displays the agent builds for a kind of card, so a collection becomes something to browse. - [Courses](https://beebox.run/docs/capabilities/courses.md): structured teaching material and per-learner progress. - [Recipes](https://beebox.run/docs/capabilities/recipes.md): scaling-aware cooking cards. - [Publishing](https://beebox.run/docs/capabilities/publishing.md): put selected box content on the public web. ## The kinds of things it holds Every card type has a reference page in [reference/cards/index.md](https://beebox.run/docs/reference/cards/index.md). - Inputs: memo, image, audio, capture-session, upload-batch, file, pdf. - Web material: webpage, extfile, commentary, tab-arrangement. - Things you keep: record, doc, person, place, recipe, inventory, figure, todo-view. - Mail and messages: email-thread, email-message, email-outbound, telegram-message, chat-thread, feedback. - Google material: gdoc, gsheet, gfolder, glink. - Teaching the agent: briefing, guide, personality, landmark, question. - Automation: procedure, procedure-run, scheduled-script, and the job cards the engine queues. - Learning: course, concept-map, exposition-plan, lesson-plan, progress. - The interface itself: dashboard, settings, browse, questions, landmarks, history, admin, view, nav. A box can also define its own card types. See [making it yours](https://beebox.run/docs/13-making-it-yours.md).