Wax Studio

Wax Studio gives you control over how Bee generates reports. Instead of getting a one-size-fits-all output, you can choose the writing style, citation format, tone, length, and which sections to include — all before generation begins.

Hive Enterprise Feature

Wax Studio is available exclusively in STING Hive (Enterprise Edition). Community Edition users receive standard report formatting.

Overview

When you request a report from Bee, Wax Studio lets you customize the output by selecting a preset or building a custom configuration. Your preferences are injected into the report generation prompt, so the LLM follows your formatting instructions from the start — not as a post-processing step.

Why “Wax”? In a beehive, beeswax is the building material that shapes the honeycomb. Wax Studio shapes the structure of your reports.

Wax Studio — Presets and customization page

How It Works

┌──────────────────────────────────────────────────────────┐
│                  Bee Chat Input Area                      │
│  ┌──────────────────────────────────┐  ┌──────────┐     │
│  │  "Write a report about..."       │  │ 🔥 Wax   │     │
│  └──────────────────────────────────┘  └──────────┘     │
└──────────────────────────────────────────────────────────┘
                         │
                         ▼
┌──────────────────────────────────────────────────────────┐
│              Wax Studio Modal Overlay                     │
│                                                          │
│  ┌─ Quick Presets ──────────────────────────────────┐   │
│  │ Executive Brief │ Research │ Technical │ Quick │  │   │
│  └──────────────────────────────────────────────────┘   │
│                                                          │
│  ┌─ Customize ──────────────────────────────────────┐   │
│  │ Style │ Citations │ Tone │ Length │ Sections      │   │
│  └──────────────────────────────────────────────────┘   │
│                                                          │
│  [ Cap & Generate ]                                      │
└──────────────────────────────────────────────────────────┘
                         │
                         ▼
┌──────────────────────────────────────────────────────────┐
│           Report Generation Pipeline                      │
│                                                          │
│  wax_config → build_wax_instructions() →                 │
│  === REPORT FORMAT INSTRUCTIONS (from Wax Studio) ===    │
│  Appended to report prompt before LLM generation         │
└──────────────────────────────────────────────────────────┘

Quick Presets

Presets are one-click configurations that set all options at once. Choose one and hit Cap & Generate to go immediately, or use a preset as a starting point and customize further.

Preset	Style	Citations	Tone	Length	Sections
Executive Brief	Concise overview	None	Formal	~1,000 words	Exec summary, no ToC/refs/appendix
Research Report	Academic with methodology	APA 7th	Analytical	~5,000 words	All sections enabled
Technical Analysis	Data-driven breakdown	IEEE	Analytical	~2,500 words	ToC, refs, appendix (no exec summary)
Quick Summary	Digestible overview	None	Conversational	~1,000 words	Minimal — no ToC, refs, or appendix
Narrative Deep Dive	Story-driven exploration	Chicago	Exploratory	~5,000 words	ToC, refs (no exec summary)

Customization Options

Report Style

Controls the overall writing approach:

Executive Brief — Concise, decision-focused, key findings up front
Research Report — Academic structure with methodology, analysis, and literature review
Technical Analysis — Data tables, implementation details, metrics-heavy
General Summary — Accessible plain-language overview
Narrative — Engaging storytelling with evidence woven in

Citation Format

Choose how sources are referenced:

APA 7th — (Author, Year) parenthetical
MLA 9th — (Author Page) in-text
Chicago — Footnotes with full reference list
IEEE — Numbered brackets [1], [2], [3]
Harvard — (Author Year, p.X) with page numbers
None — Inline attribution without formal citations

Tone / Mood

Sets the voice and register:

Analytical — Objective, evidence-based reasoning
Persuasive — Compelling arguments with supporting data
Neutral — Balanced, no editorial position
Exploratory — Open-ended, considers multiple perspectives
Formal — Professional and authoritative
Conversational — Approachable, plain language

Target Length

Guides the LLM on scope:

Concise — ~1,000 words, essential points only
Standard — ~2,500 words, balanced coverage
Comprehensive — ~5,000 words, thorough deep-dive
Exhaustive — ~10,000 words, leave no stone unturned

Section Toggles

Enable or disable structural sections:

Table of Contents — Auto-generated section navigation
Executive Summary — Top-level findings summary
References — Source citations and bibliography
Appendix — Supplementary data, tables, raw sources

Access Points

Click the flame icon in the Bee Chat input area to open the Wax Studio overlay. Select options, then Cap & Generate sends your message with the Wax configuration attached.

Wax Studio flame icon in Bee Chat input

An active Wax configuration shows an amber indicator below the chat input displaying the current preset name. Click ✕ to clear it and return to default formatting.

Wax Studio Page

Navigate to Wax Studio in the sidebar for a full-page view. Browse presets, customize settings, and save a personal default that auto-applies to all future reports.

Configuration

Feature Flag

Wax Studio is gated behind the enterprise feature flag:

# config.yml
edition:
  enterprise_features:
    wax_studio: true   # Enable Wax Studio (Hive only)

The backend returns this flag via the /api/config/features endpoint. The frontend checks isEnterpriseFeatureEnabled('waxStudio') before rendering Wax Studio UI elements.

Saved Defaults

Users can save a default Wax configuration that auto-loads when opening the modal:

Stored in localStorage under wax_default_global
Contains full config object + preset ID + timestamp
Cleared by clicking “Reset” in the modal

Pipeline Integration

Wax Studio settings flow through the entire report pipeline:

Frontend — BeeChat.jsx attaches wax_config to the API request body under context
Flask API — chatbot_routes.py passes context through to report service
Report Service — report_generators.py spreads context into the external-ai /bee/chat call
External AI — report_orchestrator.py reads request.context.get('wax_config')
Prompt Builder — build_wax_instructions() in report_prompts.py converts the config into an LLM instruction block
LLM — Receives the instruction block as part of the report prompt

The instruction block is fenced to ensure the LLM treats it as formatting rules:

=== REPORT FORMAT INSTRUCTIONS (from Wax Studio) ===
WRITING STYLE: Executive Brief — Write a concise executive summary...
CITATION FORMAT: None — Do not include a References section...
TONE: Formal — Maintain a professional, authoritative tone...
=== END REPORT FORMAT INSTRUCTIONS ===

Roadmap

Future phases of Wax Studio include:

Phase 2 — Database-backed presets (personal, team, org-level), API CRUD, default auto-apply
Phase 3 — Drag-and-drop section builder (“Wax Cells”) for custom report structure
Phase 4 — Team Foundations (admin-locked templates), Wax Molds, versioning, compliance templates

Last updated: April 12, 2026