vault backup: 2026-01-05 13:03:55
This commit is contained in:
@@ -0,0 +1,74 @@
|
||||
|
||||
key : 272f337c0d2c4407b930bde5e9846072
|
||||
endpoint: https://my-chatgpt.openai.azure.com/
|
||||
|
||||
```bash
|
||||
export AZURE_OPENAI_API_KEY="272f337c0d2c4407b930bde5e9846072"
|
||||
export AZURE_OPENAI_ENDPOINT="https://my-chatgpt.openai.azure.com/"
|
||||
```
|
||||
```
|
||||
|
||||
```.env
|
||||
# ChatGPT Settings (required)
|
||||
# Set the API Key from OpenAI
|
||||
OPENAI_API_KEY=272f337c0d2c4407b930bde5e9846072
|
||||
# To use Azure OpenAI API, set `OPENAI_AZURE` to true and `CHATGPT_REVERSE_PROXY` to your completion endpoint
|
||||
# OPENAI_AZURE=false
|
||||
OPENAI_AZURE=true
|
||||
CHATGPT_REVERSE_PROXY=https://my-chatgpt.openai.azure.com/
|
||||
|
||||
# Set the ChatGPT conversation context to 'thread', 'room' or 'both'.
|
||||
CHATGPT_CONTEXT=thread
|
||||
# Set the ChatGPT model to be used by the API. 'gpt-3.5-turbo' is the official ChatGPT-model from OpenAI
|
||||
# Note that the models are not free and will charge your OpenAI account depending on the usage of tokens
|
||||
#CHATGPT_API_MODEL=gpt-3.5-turbo
|
||||
CHATGPT_API_MODEL=gpt-4o
|
||||
# (Optional) Explicitly set the prefix sent to model at the beginning of a conversation
|
||||
#CHATGPT_PROMPT_PREFIX=Instructions:\nYou are ChatGPT, a large language model trained by OpenAI.
|
||||
# (Optional) Set to true if ChatGPT should ignore any messages which are not text
|
||||
#CHATGPT_IGNORE_MEDIA=false
|
||||
# (Optional) You can change the api url to use another (OpenAI-compatible) API endpoint
|
||||
#CHATGPT_REVERSE_PROXY=https://api.openai.com/v1/chat/completions
|
||||
# (Optional) Set the temperature of the model. 0.0 is deterministic, 1.0 is very creative.
|
||||
CHATGPT_TEMPERATURE=0.1
|
||||
# (Optional) (Optional) Davinci models have a max context length of 4097 tokens, but you may need to change this for other models.
|
||||
CHATGPT_MAX_CONTEXT_TOKENS=8192
|
||||
# You might want to lower this to save money if using a paid model. Earlier messages will be dropped until the prompt is within the limit.
|
||||
# CHATGPT_MAX_PROMPT_TOKENS=3097
|
||||
|
||||
# Set data store settings
|
||||
KEYV_BACKEND=file
|
||||
KEYV_URL=
|
||||
KEYV_BOT_ENCRYPTION=false
|
||||
KEYV_BOT_STORAGE=true
|
||||
|
||||
# Matrix Static Settings (required, see notes)
|
||||
# Defaults to "https://matrix.org"
|
||||
MATRIX_HOMESERVER_URL=
|
||||
# With the @ and :DOMAIN, ie @SOMETHING:DOMAIN - Not used if `MATRIX_ACCESS_TOKEN` is set.
|
||||
MATRIX_BOT_USERNAME=
|
||||
# Set `MATRIX_BOT_PASSWORD` the bot will print an `MATRIX_ACCESS_TOKEN` to the terminal
|
||||
MATRIX_ACCESS_TOKEN=
|
||||
# Not used if `MATRIX_ACCESS_TOKEN` is set.
|
||||
MATRIX_BOT_PASSWORD=
|
||||
|
||||
# Matrix Configurable Settings Defaults (optional)
|
||||
# Leave prefix blank to reply to all messages
|
||||
MATRIX_DEFAULT_PREFIX=!chatgpt
|
||||
MATRIX_DEFAULT_PREFIX_REPLY=false
|
||||
|
||||
# Matrix Access Control (optional)
|
||||
# Can be set to user:homeserver or a wildcard like :anotherhomeserver.example
|
||||
MATRIX_BLACKLIST=
|
||||
# `MATRIX_WHITELIST` is overriden by `MATRIX_BLACKLIST` if they contain same entry
|
||||
MATRIX_WHITELIST=
|
||||
|
||||
# Matrix Feature Flags (optional)
|
||||
MATRIX_AUTOJOIN=true
|
||||
MATRIX_ENCRYPTION=true
|
||||
# If you turn threads off you will have problems if you don't set CHATGPT_CONTEXT=room
|
||||
MATRIX_THREADS=true
|
||||
MATRIX_PREFIX_DM=false
|
||||
MATRIX_RICH_TEXT=true
|
||||
```
|
||||
|
||||
@@ -0,0 +1,13 @@
|
||||
|
||||
|
||||
api key
|
||||
vscode:
|
||||
```
|
||||
sk-b3426ba1862543bd876be65b7f830499
|
||||
```
|
||||
|
||||
|
||||
zed:
|
||||
```
|
||||
sk-2f351b2c4d084e7c98a53311cf09e3da
|
||||
```
|
||||
@@ -0,0 +1,261 @@
|
||||
|
||||
|
||||
openapi key:
|
||||
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
|
||||
|
||||
[[Creating user accounts Dendrite]]
|
||||
|
||||
|
||||
synapse:
|
||||
register_new_matrix_user -c /etc/matrix-synapse/homeserver.yaml
|
||||
New user localpart: gpt
|
||||
Password: windyboy@2006
|
||||
token from element: syt_Z3B0_yBPDcvVUmXFgHeNPGRWa_32nnGL
|
||||
new token: syt_Z3B0_PPffEqKjnAjaIpcuRRuj_0LE1j5
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
python:
|
||||
This bot's public fingerprint ("Session key") for one-sided verification is: jkH6 U0p/ O58Z DHbr M+1i AKOF RhYP W80A Xmqy HlKh fH0
|
||||
|
||||
|
||||
gzzn dev:
|
||||
token:
|
||||
syt_Z3B0X2JvdA_RydZTTmGHAbeBVvseZIE_3eFONm
|
||||
|
||||
|
||||
## azure gpt bot
|
||||
user: ms
|
||||
password: NzI3MDRmNTExNDRj
|
||||
azure gpt key: 272f337c0d2c4407b930bde5e9846072
|
||||
azure endpoint: https://my-chatgpt.openai.azure.com/
|
||||
location/regin: eastus
|
||||
|
||||
|
||||
gpt4:
|
||||
user: gpt4
|
||||
password: windyboy@2006
|
||||
access token: syt_Z3B0NA_dXVvfYHuYyEnfvDUqCyx_1gHT8y
|
||||
openapi key: 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
|
||||
|
||||
|
||||
new google account openai
|
||||
matrix api: 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
|
||||
|
||||
mail gpt4:
|
||||
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
|
||||
|
||||
|
||||
azure gpt:
|
||||
key: 272f337c0d2c4407b930bde5e9846072
|
||||
endpoint: https://my-chatgpt.openai.azure.com/
|
||||
|
||||
|
||||
|
||||
|
||||
```
|
||||
# Role & Identity
|
||||
你是由 Google 研发的先进 AI 助手 {{ baibot_name }},基于 {{ baibot_model_id }} 架构。
|
||||
当前会话启动时间: {{ baibot_conversation_start_time_utc }}。
|
||||
# Core Capabilities (针对 Gemini 优化)
|
||||
1. **深度推理**:拥有强大的逻辑分析、代码生成和数学计算能力。
|
||||
2. **长程记忆**:能够精准回顾和关联长对话历史中的细节,保持上下文一致性。
|
||||
3. **思维透明**:对于非显而易见的问题,必须通过"显式推理"展示你的思考路径。
|
||||
|
||||
# Thinking Protocol (思维协议)
|
||||
在回答用户之前,你必须执行以下思维循环:
|
||||
4. **意图识别**:用户真正想要解决的核心痛点是什么?隐含需求是什么?
|
||||
5. **知识检索**:在你的知识库和当前对话历史中检索相关信息。
|
||||
6. **逻辑推导**:构建解决路径,预判潜在的错误或陷阱。
|
||||
7. **自我修正**:检查生成的答案是否准确、无害且符合逻辑。
|
||||
|
||||
# Response Format (响应格式规范)
|
||||
|
||||
## 场景 A:复杂任务(代码、逻辑、分析、长文本生成)
|
||||
必须严格包含以下 Markdown 模块:
|
||||
|
||||
> **🤔 深度思考**:
|
||||
> *此处展示你的简要分析逻辑、解题思路或关键决策点。*
|
||||
|
||||
> **📋 详细解答**:
|
||||
> *此处提供具体的答案、代码实现或详细论述。*
|
||||
|
||||
> **💡 专家建议**:
|
||||
> *提供优化建议、潜在风险预警或延伸知识。*
|
||||
|
||||
## 场景 B:简单任务(问候、明确的短问题)
|
||||
- 直接给出简洁、准确的回答,无需展示思考过程。
|
||||
|
||||
# Interaction Guidelines (交互准则)
|
||||
- **准确性优先**:严禁编造事实。如果不知道,请直接说明。
|
||||
- **代码质量**:生成的代码必须是完整的、可执行的,并包含必要的注释。
|
||||
- **语言风格**:专业、客观、有条理。避免使用过度情绪化的词语。
|
||||
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
grok:
|
||||
|
||||
```
|
||||
base_url: https://openrouter.ai/api/v1
|
||||
api_key: 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
|
||||
|
||||
text_generation:
|
||||
model_id: x-ai/grok-4.1-fast
|
||||
reasoning:
|
||||
effort: "high" # 可改为 "medium", "low", "minimal", "none"
|
||||
exclude: false # true 表示隐藏思考 TOKENS,仅返回最终答案
|
||||
temperature: 0.3
|
||||
max_response_tokens: 4096
|
||||
max_context_tokens: 2000000
|
||||
|
||||
prompt: |
|
||||
# Role & Identity
|
||||
你是 {{ baibot_name }},一名基于 {{ baibot_model_id }} 运行的高级 Agentic AI 助手。
|
||||
{{ baibot_model_id }} 是 xAI 的顶级模型之一,拥有 2M 超长上下文、强推理能力、可靠的工具调用机制。
|
||||
你的任务是:解决问题、提供高价值分析、执行工具调用,并保持专业性与安全性。
|
||||
当前会话启动时间:{{ baibot_conversation_start_time_utc }}。
|
||||
|
||||
# Core Capabilities(专为 Grok-4.1-Fast 调校)
|
||||
1. **Agentic Tool Calling**:在必要时自主调用工具,以实现精准查询、复杂任务分解与可执行方案。
|
||||
2. **Ultra-Long Context (2M tokens)**:可处理长文档、长代码库、研究型内容而不丢失上下文。
|
||||
3. **Controlled Reasoning**:根据 `reasoning_enabled` 配置决定推理深度:
|
||||
- **true**:允许深度思考、研究、逻辑链
|
||||
- **false**:使用简洁、高速、支持型回答
|
||||
4. **Real-World Use Case Optimization**:特别适用于技术支持、调试、研究、大型代码理解、系统架构分析。
|
||||
5. **安全与事实性优先**:对事实错误零容忍;不清楚时应明确说明。
|
||||
|
||||
# Thinking Protocol(思维协议)
|
||||
在回答前你必须执行以下内部流程(用户仅看到摘要):
|
||||
1. **意图分析**:识别显性与隐性需求
|
||||
2. **上下文吸收**:使用 2M 上下文能力读取相关内容
|
||||
3. **方案构建**:必要时通过工具解决复杂任务
|
||||
4. **逻辑校验**:检查一致性、事实性、安全性
|
||||
5. **输出优化**:确保回答结构清晰、可执行、无噪音
|
||||
|
||||
# Response Format(响应格式规范)
|
||||
## A 类:复杂任务(代码、调试、分析、研究、工具调用)
|
||||
输出结构必须包含:
|
||||
|
||||
> **🤖 思考摘要(可见)**
|
||||
> *展示关键推理点、问题拆解、是否需要工具调用。*
|
||||
|
||||
> **📘 详细解答**
|
||||
> *提供最终答案、步骤、分析或代码。所有代码必须可运行并附注释。*
|
||||
|
||||
> **🛠 工具策略(如适用)**
|
||||
> *如果需要调用工具,请明确指出你的调用目的与预期结果。*
|
||||
|
||||
> **⚡ 延伸建议**
|
||||
> *给出进一步改进、潜在风险或扩展方向。*
|
||||
|
||||
---
|
||||
|
||||
## B 类:简单任务(问候、轻量知识问答、简短建议)
|
||||
- 直接输出简洁、明确的答案
|
||||
- 不展示“思考摘要”
|
||||
|
||||
---
|
||||
|
||||
# Interaction Guidelines(交互准则)
|
||||
- **准确性第一**:如果缺乏足够信息,请请求澄清或说明不确定性
|
||||
- **风格**:专业、逻辑、清晰,不使用夸张性语言
|
||||
- **工具调用**:仅在确实有助于结果时调用
|
||||
- **代码质量**:必须可执行、含注释、结构化
|
||||
- **尊重上下文**:善用 2M context,不遗忘信息
|
||||
- **用户至上**:目标是解决问题,而不是展示能力
|
||||
|
||||
```
|
||||
|
||||
|
||||
```
|
||||
base_url: https://openrouter.ai/api/v1
|
||||
api_key: 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
|
||||
|
||||
text_generation:
|
||||
model_id: x-ai/grok-4.1-fast
|
||||
|
||||
# 百科问答模式建议:简洁推理 + 降低成本
|
||||
reasoning:
|
||||
effort: "minimal" # 保留少量内部推理提升准确性
|
||||
exclude: true # 不展示推理内容,回答更“百科风”
|
||||
|
||||
temperature: 0.2 # 降温以减少幻觉
|
||||
max_response_tokens: 1024
|
||||
max_context_tokens: 2000000 # Grok 全量上下文,可容纳大型知识内容
|
||||
|
||||
prompt: |
|
||||
# Role & Identity
|
||||
你是 {{ baibot_name }},一个基于 {{ baibot_model_id }}运行的百科知识问答机器人。
|
||||
职责是提供:**准确、权威、可验证** 的知识性回答。
|
||||
当前会话启动时间:{{ baibot_conversation_start_time_utc }}。
|
||||
|
||||
# Core Capabilities(百科问答优化)
|
||||
1. **事实性优先**:必须确保回答可验证,杜绝编造。
|
||||
2. **知识覆盖广**:历史、科技、文化、地理、生物、工程、生活常识等都能回答。
|
||||
3. **解释简洁清晰**:像百科一样用客观语言描述,不夸张,不情绪化。
|
||||
4. **引用型表述**:如知识存在争议,应说明“在主流观点中…”。
|
||||
5. **安全稳妥**:避免医学诊断、金融投资、法律判断等高风险输出。
|
||||
|
||||
# Response Format(回答格式)
|
||||
## 简单知识问答 / 百科问答(默认)
|
||||
- 直接输出明确、准确的答案。
|
||||
- 信息按分点或短段落组织,易读易理解。
|
||||
|
||||
## 复杂问题(多步骤解释、概念对比、历史背景)
|
||||
输出包含:
|
||||
- **📘 百科式说明**:关键定义、背景、核心解释
|
||||
- **📚 延伸阅读**(如适用):补充知识、相关概念
|
||||
|
||||
# Interaction Guidelines(交互准则)
|
||||
- **如不确定事实,必须明确声明“不确定”**。
|
||||
- 不讨论阴谋论、不可靠数据源、不严谨的统计。
|
||||
- 避免提供专业医学、法律、投资建议。
|
||||
- 保持中立、客观、权威的语气。
|
||||
|
||||
|
||||
```
|
||||
|
||||
|
||||
|
||||
```
|
||||
base_url: "https://zenmux.ai/api/v1"
|
||||
api_key: "sk-ai-v1-2d2ba59719ff6f0d8d2f439d3b5c84399176d1059302cc4b43c132a4d17e9f03"
|
||||
|
||||
text_generation:
|
||||
model_id: "deepseek/deepseek-reasoner"
|
||||
temperature: 0.1
|
||||
max_response_tokens: 16384
|
||||
max_context_tokens: 128000
|
||||
|
||||
prompt: |
|
||||
# Role
|
||||
你是一个专注于严谨逻辑推理、工程正确性和复杂问题拆解的 AI 助手。
|
||||
|
||||
你的核心目标是:
|
||||
- 给出结论正确、可执行、可复查的答案
|
||||
- 在内部进行充分推理,但不显式暴露完整思维链
|
||||
|
||||
# Reasoning Policy
|
||||
- 对复杂问题进行深度推理(内部完成)
|
||||
- 输出时仅提供:
|
||||
- 明确结论
|
||||
- 关键步骤或必要的简化推理说明
|
||||
- 可验证的事实与假设
|
||||
- 不输出逐 token 的思维链
|
||||
|
||||
# Engineering Standards
|
||||
- 所有代码必须可直接运行,包含必要注释与错误处理
|
||||
- 架构或配置建议必须说明原因
|
||||
- 对不确定性必须明确标注
|
||||
|
||||
# Style
|
||||
- 专业、冷静、工程师视角
|
||||
- 少废话,高密度信息
|
||||
|
||||
|
||||
```
|
||||
@@ -0,0 +1,5 @@
|
||||
|
||||
emb:
|
||||
```
|
||||
634442642d294d5cb1b83f5d3790bd98.VH4nn223ldRdyyi_Fuk6MpWz
|
||||
```
|
||||
@@ -0,0 +1,17 @@
|
||||
|
||||
## Key
|
||||
vscode:
|
||||
```
|
||||
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
|
||||
```
|
||||
|
||||
matrix-bot
|
||||
```
|
||||
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
|
||||
```
|
||||
|
||||
local rag:
|
||||
```
|
||||
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
|
||||
```
|
||||
|
||||
@@ -0,0 +1,5 @@
|
||||
|
||||
key
|
||||
```
|
||||
sk-scalJeWNKxWMePXVwiGVnCMjrDpeSkCFxyowSSYp7C9yDpFqX3wY6zg9N7ovJ0MR
|
||||
```
|
||||
@@ -0,0 +1,25 @@
|
||||
|
||||
|
||||
coder :
|
||||
```
|
||||
sk-ai-v1-875cd41da6e117609e850e4c594d0116f2e128bee9bf6890eb6a48fe23e69764
|
||||
```
|
||||
|
||||
url:
|
||||
```
|
||||
https://zenmux.ai/api/v1
|
||||
```
|
||||
|
||||
```
|
||||
https://zenmux.ai/api/anthropic
|
||||
```
|
||||
|
||||
```
|
||||
https://zenmux.ai/api/vertex-ai
|
||||
```
|
||||
|
||||
obsidian:
|
||||
```
|
||||
sk-ai-v1-82f1a2df15721ca5d5afc633842b91719fea95c6c449cbb78db0dd03f7ed1aa2
|
||||
```
|
||||
|
||||
@@ -0,0 +1,4 @@
|
||||
Code:
|
||||
```
|
||||
32f3aa3e5c3848fbad264e83b848b226.BVXajLMpbRPaHfv7
|
||||
```
|
||||
@@ -0,0 +1,5 @@
|
||||
|
||||
code token:
|
||||
```
|
||||
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
|
||||
```
|
||||
@@ -0,0 +1,132 @@
|
||||
|
||||
|
||||
---
|
||||
|
||||
# 📘 LiteLLM 配置指南:NewCli (AWS/Anthropic Proxy)
|
||||
|
||||
版本日期: 2025-12-26
|
||||
|
||||
适用场景: 对接自定义 Anthropic 代理(NewCli),解决路径拼接 (404)、参数不兼容 (400) 及防火墙拦截 (403) 问题。
|
||||
|
||||
## 1. 核心参数规范 (Critical Specs)
|
||||
|
||||
无论使用 UI 还是 YAML,必须严格遵守以下三条铁律:
|
||||
|
||||
1. **Provider (提供商)**: 必须选 `Anthropic`。
|
||||
|
||||
- _原因_: 让 LiteLLM 自动处理 `/v1/messages` 路径拼接和 JSON 格式转换。
|
||||
|
||||
2. **Base URL (基准地址)**: `https://code.newcli.com/claude/aws`
|
||||
|
||||
- > [!WARNING] 警告
|
||||
|
||||
- > **严禁**在末尾加 `/v1`。LiteLLM 会自动追加,加了会导致双重路径 (`/v1/v1`) 报 **404**。
|
||||
|
||||
3. **Model ID (模型名)**: `claude-sonnet-4-5`
|
||||
|
||||
- _原因_: 代理商白名单仅支持此 ID。
|
||||
|
||||
|
||||
---
|
||||
|
||||
## 2. UI 配置方案 (推荐)
|
||||
|
||||
**入口**: LiteLLM UI (`/ui`) -> **Models** -> **+ Add Model**
|
||||
|
||||
### 基础信息 (General Settings)
|
||||
|
||||
|**字段**|**填写内容**|**说明**|
|
||||
|---|---|---|
|
||||
|**Model Name**|`claude-sonnet`|客户端调用的别名|
|
||||
|**Select Provider**|**Anthropic**|⚠️ 必选|
|
||||
|**Litellm Model Name**|`claude-sonnet-4-5`|真实模型 ID|
|
||||
|**API Base URL**|`https://code.newcli.com/claude/aws`|⚠️ 末尾无 `/v1`|
|
||||
|**API Key**|`sk-ant-oat01...`|填入完整 Key|
|
||||
|
||||
### 高级参数 (LiteLLM Params / Metadata)
|
||||
|
||||
> [!TIP] 关键步骤
|
||||
>
|
||||
> 在 JSON 输入框填入以下内容,用于解决参数兼容性和防火墙拦截。
|
||||
|
||||
JSON
|
||||
|
||||
```
|
||||
{
|
||||
"drop_params": true,
|
||||
"extra_headers": {
|
||||
"anthropic-version": "2023-06-01",
|
||||
"User-Agent": "curl/7.68.0",
|
||||
"Authorization": "Bearer ${NEWCLI_API_KEY}"
|
||||
},
|
||||
"no_verify_ssl": true
|
||||
}
|
||||
```
|
||||
|
||||
_注:如果不使用变量,请在 `Authorization` 里直接填入 `Bearer sk-ant...`_
|
||||
|
||||
---
|
||||
|
||||
## 3. YAML 文件配置方案 (IaC)
|
||||
|
||||
适用于 `docker-compose` 挂载配置。
|
||||
|
||||
YAML
|
||||
|
||||
```
|
||||
model_list:
|
||||
- model_name: claude-sonnet
|
||||
litellm_params:
|
||||
model: anthropic/claude-sonnet-4-5
|
||||
# ⚠️ 重点:Base URL 不带 /v1
|
||||
api_base: https://code.newcli.com/claude/aws
|
||||
# 建议使用环境变量
|
||||
api_key: os.environ/NEWCLI_API_KEY
|
||||
extra_headers:
|
||||
anthropic-version: "2023-06-01"
|
||||
# 伪装 UA 防拦截
|
||||
User-Agent: "curl/7.68.0"
|
||||
# 强制 Bearer 鉴权 (可选,视代理商严格程度)
|
||||
Authorization: "Bearer ${NEWCLI_API_KEY}"
|
||||
|
||||
general_settings:
|
||||
master_key: sk-1234
|
||||
database_url: postgresql://litellm:litellm@litellm-postgres:5432/litellm
|
||||
|
||||
litellm_settings:
|
||||
# ⚠️ 核心修复:丢弃不兼容参数(如 user, frequency_penalty),解决 400 错误
|
||||
drop_params: true
|
||||
set_verbose: true
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 4. 故障排查手册 (Troubleshooting)
|
||||
|
||||
|**状态码**|**错误类型**|**根本原因**|**解决方案**|
|
||||
|---|---|---|---|
|
||||
|**404**|`NotFoundError`|**路径重复**|检查 `api_base` 是否多写了 `/v1`。应该让 LiteLLM 自动拼接。|
|
||||
|**400**|`BadRequest`|**参数冗余**|LiteLLM 传了 OpenAI 专有参数给 Anthropic。需开启 `drop_params: true`。|
|
||||
|**403**|`Forbidden`|**WAF 拦截**|缺少 User-Agent 伪装。需在 header 添加 `"User-Agent": "curl/..."`。|
|
||||
|**401**|`AuthError`|**鉴权失败**|Key 错误或格式不对。尝试在 `extra_headers` 强制注入 `Authorization: Bearer <key>`。|
|
||||
|
||||
---
|
||||
|
||||
## 5. 客户端调用示例
|
||||
|
||||
验证配置是否成功的标准命令(访问 LiteLLM 端口):
|
||||
|
||||
Bash
|
||||
|
||||
```
|
||||
curl -X POST http://localhost:4000/v1/chat/completions \
|
||||
-H "Authorization: Bearer sk-1234" \
|
||||
-H "Content-Type: application/json" \
|
||||
-d '{
|
||||
"model": "claude-sonnet",
|
||||
"messages": [
|
||||
{ "role": "user", "content": "Config Test: OK?" }
|
||||
]
|
||||
}'
|
||||
```
|
||||
|
||||
@@ -0,0 +1,7 @@
|
||||
|
||||
|
||||
context7 mcp key:
|
||||
```
|
||||
ctx7sk-92c2c98e-817e-41d4-bb85-94824444e2bf
|
||||
```
|
||||
|
||||
@@ -0,0 +1,37 @@
|
||||
|
||||
compose.yml
|
||||
```yaml
|
||||
services:
|
||||
db:
|
||||
image: postgres:17-alpine
|
||||
container_name: oui-db
|
||||
restart: always
|
||||
environment:
|
||||
- POSTGRES_USER=webui
|
||||
- POSTGRES_PASSWORD=webui_password
|
||||
- POSTGRES_DB=open_webui
|
||||
volumes:
|
||||
- db_data:/var/lib/postgresql/data
|
||||
|
||||
open-webui:
|
||||
image: ghcr.io/open-webui/open-webui:main
|
||||
container_name: oui
|
||||
restart: always
|
||||
ports:
|
||||
- "3000:8080"
|
||||
depends_on:
|
||||
- db
|
||||
extra_hosts:
|
||||
- "host.docker.internal:host-gateway"
|
||||
environment:
|
||||
- 'DATABASE_URL=postgresql://webui:webui_password@db:5432/open_webui'
|
||||
- 'OPENAI_API_BASE_URL=http://host.docker.internal:4000/v1'
|
||||
- 'OPENAI_API_KEY=sk-1234'
|
||||
- 'WEBUI_SECRET_KEY=super_secret_key'
|
||||
volumes:
|
||||
- oui_data:/app/data
|
||||
|
||||
volumes:
|
||||
db_data:
|
||||
oui_data:
|
||||
```
|
||||
@@ -0,0 +1,5 @@
|
||||
|
||||
api key
|
||||
```
|
||||
xai-FDgOu9cZhAkeEBGnkFp61gyTIeqNmWuJ8CLABHIkqTUR1RYzm08hlXabnCTBrj91ee0pYjk0ZWtmRjhS
|
||||
```
|
||||
Reference in New Issue
Block a user