v0.5.2

add: The channel icon to which the model belongs
add: aws Cohere
Fix: aws Claude Original Response
add: Clear channel statistics
feat: Optimize claude detailed logging
Fix: the problem of automatically retrying streaming data in the openai channel
Fix Claude oai non-streaming format
feat: 思考过程劫持功能（#108）
feat: add Claude thinking (#111)
fix: Optimize audio interface processing logic to distinguish between TTS and Whisper request formats
Fix: Aws Repeated Try Question
feat: optimize audio token calculation and log output

v0.5.1

Feat: Add support for cache_control for the Claude channel.
Feat: Update aws sdk
Feat: add deepseek r1
feat: add support for o3-mini models in model ratio
Fix: Whisper-1 pricing
Style: User settings page update
Fix: Failed to retrieve images in Midjourney
Fix: Search results display for the Users group
Add: OpenAI real-time models are now supported

v0.5.0

Fix: The tool for aws, gcp, and Claude's /v1/messages endpoint.
add: login log and error return log
feat:OpenAI's O1 is stream billing.
add: Channel Tag Management
fix: Optimize the bug that 502 may appear when opening the setting page
feat: Testing if the channel returns the correct information.
add: gcp claude-20241022

v0.4.9

feat: Supports user-defined spending limits and alerts
Add: gcp claude-3-5-sonnet-20241022 model
Fix: AWS is showing the correct status code
fix: 重定向模型无法测试 #88
fix: 渠道测试的弹窗滚动优化 （#87）
fix: Optimize the display of the conversation page

v0.4.8

Add: Support conversation connection jump
feat: Optimization announcement displayed in the show
feat: update model ratio
Fix: Optimize Relay
feat: format o1 model max tokens param
Fix: claude tool_choice may only be specified while providing tools #82
Fix: 日志显示调整 #81
feat: support OpenAI o1-preview and o1-mini
Fix the bug that claude official request format cannot recognize images
Add Handling 529 overloads for GCP claude flow-throughs
Add gemini 0827 model
Optimize the SMS verification code sent during registration (#75 #67)

v0.4.7

Add support for the backend to get channel models
Channel support allows you to add multiple models separated by commas.
Add a button that triggers a pop-up announcement.
Optimize model billing acquisition
Add chatgpt-4o-latest model

v0.4.6

Fix the determination of empty responses in the speech model
Fix: Increase in user quota and recharge quota
Add IP display to request logs
Add chart statistics display for client-side logsgit
Add date selection for the customer-facing homepage
feat: support gpt-4o mini
Add the ability for C-end users to set token groups
Fix: Claude channel opening to support original requests

v0.4.5

支持GCP-Claude，支持rt、密钥两种方式
支持单渠道设置代理请求
优化gpt-4o计费
优化Claude、讯飞中断请求扣费为1token

v0.4.4

支持新增渠道状态码复写
claude支持System
C端支持深色模式
feat: support user content download proxy & relay proxy now
feat: support gemini embeddings (text-embedding-004,embedding-001)

v0.4.3

新增豆包、DeepSeek、DeepL、togetherai、Gemini1.5
优化了重试
优化管理端显示
C端增加深色模式
支持gpt-4o