Skip to main content

AI Assistant Overview

BeeCount integrates AI features to make recording smarter and easier.

Features

FeatureDescription
Voice RecordingRecord with just a sentence
Image RecognitionTake a photo to auto-recognize bills
AI ChatSmart financial assistant

Supported AI Services

  • Zhipu GLM - Available in China, recommended
  • Custom Providers - Any OpenAI-compatible provider
  • Local Model - Offline capable (in development)

Model Selection

Zhipu GLM offers multiple models to choose from:

ModelFeaturesRecommended For
GLM-4-FlashFree, fastDaily recording, recommended
GLM-4-AirBalanced performance and costBetter comprehension needed
GLM-4-AirXStronger reasoningComplex scenarios
GLM-4-FlashXEnhanced free modelFree but better results needed
GLM-4-PlusFlagship modelBest performance
GLM-4-LongUltra-long contextLong conversation scenarios

Recommended: Use GLM-4-Flash - it's free and sufficient for daily recording needs.

Configure AI

  1. Go to "Me" → "Smart Recording" → "AI Assistant", or "Discover" → "Common Features" → "AI Settings"
  2. Select AI service provider
  3. Enter API Key
  4. Save configuration

Custom AI Providers

In addition to the built-in Zhipu GLM, you can add any OpenAI-compatible provider:

Supported Providers

  • SiliconFlow - Available in China, affordable pricing
  • DeepSeek - Chinese LLM, great value
  • OpenAI - Requires overseas network access
  • Other Compatible Services - Any service providing OpenAI-compatible API

Add Custom Provider

  1. Go to "Me" → "Smart Recording" → "AI Assistant"
  2. Tap "Add Provider"
  3. Fill in provider details:
    • Name - Custom name for easy identification
    • API URL - API endpoint provided by the service
    • API Key - Your API key
    • Model Names - Configure text/vision/speech models as needed

Capability Binding

You can bind different providers to different AI capabilities:

CapabilityDescriptionExample
Text UnderstandingVoice-to-bill, AI chatUse DeepSeek
Image RecognitionPhoto bill recognitionUse Zhipu GLM
Speech RecognitionVoice to textUse Zhipu GLM

This allows you to leverage each provider's strengths for the best experience.

Get API Key

Zhipu AI

  1. Visit Zhipu Open Platform
  2. Register an account
  3. Go to "API Keys" page and create a new API Key
  4. Copy the API Key to BeeCount

💡 New users typically have free credits available after registration. GLM-4-Flash model is completely free.

SiliconFlow

  1. Visit SiliconFlow
  2. Register and complete identity verification
  3. Get API Key from the console
  4. API URL: https://api.siliconflow.cn/v1

DeepSeek

  1. Visit DeepSeek Platform
  2. Register an account
  3. Create an API Key
  4. API URL: https://api.deepseek.com/v1

Web AI Configuration

Sign in to BeeCount Cloud on the web → avatar menu → AI Config:

  • Provider CRUD — add, edit, or delete custom providers. The built-in "Zhipu GLM" cannot be deleted (matches mobile constraint); deleting a provider that's currently bound to a capability automatically falls back to the built-in.
  • Capability binding — three Select dropdowns (text chat / vision / speech), one per capability; providers that lack the relevant model are disabled.
  • Test buttons — every model input has a [Test] button on the right that runs a real LLM call through the server (CORS-free); a "Test all" button at the bottom runs all configured capabilities serially and shows server-returned errors inline on failure.
  • Cross-device sync — web changes push to mobile in real time via profile_change WS, and vice versa.

API keys are shown in plain text by default (matching mobile); be mindful when sharing your screen. Deep parameters (custom_prompt / strategy / bill_extraction_enabled / use_vision) remain on mobile and are not exposed on the web.