Files
hunjing/shenshi/config/models.example.yaml
chenxin 462d570751 M4b:ai_provider 删除手术①-⑥——神识全面接管模型出口
- ① 后端删除 app/ai_provider/(38 文件,含 anthropic/openai 格式、billing、audio、embedding)、services/ai_provider_admin.py、core/config/model_config.py、tests/ai_provider/(21 文件)
- ② 连带退役 application/agent/(旧 Agent 循环/steer/follow-up/确认注册表)、application/skill_executor.py、app/skills/(python skills 孤儿)、endpoints/audio.py、application/imports.py(ProviderManager 消费方);settings 端点收敛为生成偏好
- ③ runs.py 修剪:AgentRunRegistry/steer/follow-up/confirmations 端点退役,保留 list/get/events/cancel/stream;safety 确认改为直接落库裁决(pending→approved/denied)
- ④ 前端删除:models 页、ModelGatewayView、gateway.ts、GatewayPanel(玄关)、ModelGatewayPanel/ModelStrategyForm、audio.ts/useVoiceInput、网关类型与 mock;设置页收敛为外观/语言/通知
- ⑤ 配置收敛:config/models|providers|pricing.example.yaml 归档 shenshi/config/;.env.example 删 provider keys/MODEL_ENDPOINT_POOL/MODEL_CONFIG_PATH,收敛 SHENSHI_*
- ⑥ 验收:rg 审计灵台 Provider SDK import 与 ai_provider 引用归零;后端 528 passed(2 个预存失败经 HEAD 复跑验证与本手术无关)、前端 type-check+201 passed、神识 go vet+test、BFF 4 passed
2026-08-27 10:02:44 +08:00

240 lines
6.0 KiB
YAML
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# 魂镜 AI Provider 模型配置
# 提供商基础设施配置移到了 config/providers.yaml
version: 1
models:
# ==================== DeepSeek ====================
- name: deepseek-v4-flash
display_name: DeepSeek V4 Flash
provider: deepseek
model: deepseek-v4-flash
format: openai
tags: [default]
supports_tools: true
supports_structured_output: true
supports_streaming: true
supports_thinking: true
supports_reasoning_effort: true
context_window: 1000000
response_reserve: 8192
max_output_tokens: 384000
timeout_seconds: 300
routing_strategy: priority_failover
endpoint_pool_enabled: true
endpoints:
- name: primary
model: deepseek-v4-flash
priority: 10
weight: 100
rpm_limit: 2500
- name: fallback
model: deepseek-v4-pro
priority: 20
weight: 50
- name: deepseek-v4-pro
display_name: DeepSeek V4 Pro
provider: deepseek
model: deepseek-v4-pro
format: openai
tags: [fallback]
supports_tools: true
supports_structured_output: true
supports_streaming: true
supports_thinking: true
supports_reasoning_effort: true
context_window: 1000000
response_reserve: 8192
max_output_tokens: 384000
timeout_seconds: 300
routing_strategy: priority_failover
endpoint_pool_enabled: true
endpoints:
- name: primary
model: deepseek-v4-pro
priority: 10
weight: 100
- name: deepseek-v4-pro-anthropic
display_name: DeepSeek Pro(Anthropic 格式)
provider: deepseek
model: deepseek-v4-pro
format: anthropic
supports_tools: true
supports_structured_output: true
supports_streaming: true
supports_thinking: true
supports_reasoning_effort: true
context_window: 1000000
response_reserve: 8192
max_output_tokens: 384000
timeout_seconds: 300
endpoints:
- name: primary
model: deepseek-v4-pro
base_url_env: DEEPSEEK_ANTHROPIC_BASE_URL
priority: 10
weight: 100
# ==================== Anthropic / Claude ====================
- name: claude-sonnet
display_name: Claude Sonnet
provider: anthropic
model: claude-sonnet-4-6
format: anthropic
supports_tools: true
supports_structured_output: true
supports_streaming: true
supports_thinking: true
supports_reasoning_effort: true
supports_vision: true
context_window: 1000000
response_reserve: 16000
max_output_tokens: 16000
timeout_seconds: 60
reasoning_effort: medium
- name: claude-haiku
display_name: Claude Haiku
provider: anthropic
model: claude-haiku-4-5
format: anthropic
supports_tools: true
supports_structured_output: true
supports_streaming: true
supports_vision: true
context_window: 200000
response_reserve: 4096
max_output_tokens: 4096
timeout_seconds: 60
- name: claude-fable
display_name: Claude Fable 5
provider: anthropic
model: claude-fable-5
format: anthropic
supports_tools: true
supports_structured_output: true
supports_streaming: true
supports_thinking: true
supports_reasoning_effort: true
supports_vision: true
context_window: 1000000
response_reserve: 16000
max_output_tokens: 16000
timeout_seconds: 120
reasoning_effort: high
# ==================== Embedding ====================
# 本地元炁推理引擎(默认,零配置可用)
- name: bge-m3
display_name: BGE-M3 Embedding(元炁本地)
provider: yuanqi
model: bge-m3
format: openai
tags: [embedding]
timeout_seconds: 30
dimensions: 1024
endpoints:
- name: primary
model: bge-m3
base_url_env: YUANQI_BASE_URL
api_key_env: YUANQI_API_KEY
priority: 10
weight: 100
# 云端备选(配好 API_KEY 后改为默认)
# - name: text-embedding-3-small
# display_name: Text Embedding 3 Small
# provider: deepseek
# model: text-embedding-3-small
# format: openai
# tags: [embedding]
# timeout_seconds: 60
# dimensions: 1536
# ==================== 小米 MiMo ====================
- name: mimo-v2.5-pro
display_name: MiMo v2.5 Pro
provider: xiaomi
model: mimo-v2.5-pro
format: openai
supports_tools: true
supports_structured_output: true
supports_streaming: true
supports_reasoning_effort: true
supports_vision: true
context_window: 128000
response_reserve: 8192
max_output_tokens: 32768
timeout_seconds: 300
routing_strategy: priority_failover
endpoint_pool_enabled: true
endpoints:
- name: primary
model: mimo-v2.5-pro
priority: 10
weight: 100
- name: mimo-v2.5
display_name: MiMo v2.5
provider: xiaomi
model: mimo-v2.5
format: openai
supports_tools: true
supports_structured_output: true
supports_streaming: true
supports_vision: true
context_window: 128000
response_reserve: 4096
max_output_tokens: 16384
timeout_seconds: 180
- name: mimo-v2.5-asr
display_name: MiMo v2.5 ASR
provider: xiaomi
model: mimo-v2.5-asr
format: openai
tags: [asr]
context_window: 0
timeout_seconds: 120
endpoints:
- name: primary
model: mimo-v2.5-asr
priority: 10
weight: 100
- name: mimo-v2.5-tts
display_name: MiMo v2.5 TTS
provider: xiaomi
model: mimo-v2.5-tts
format: openai
tags: [tts]
supports_streaming: true
context_window: 0
timeout_seconds: 120
- name: mimo-v2.5-tts-voiceclone
display_name: MiMo v2.5 TTS 声纹克隆
provider: xiaomi
model: mimo-v2.5-tts-voiceclone
format: openai
tags: [tts, voice_clone]
supports_streaming: true
context_window: 0
timeout_seconds: 300
- name: mimo-v2.5-tts-voicedesign
display_name: MiMo v2.5 TTS 音色设计
provider: xiaomi
model: mimo-v2.5-tts-voicedesign
format: openai
tags: [tts, voice_design]
supports_streaming: true
context_window: 0
timeout_seconds: 300