models-config

Pi Models Manager

Manage ~/.pi/agent/models.json command-line tool for configuration files.

Quick Start

# 列出所有模型
bun ~/.pi/agent/skills/models-config/models.ts list

# 添加 Provider
bun ~/.pi/agent/skills/models-config/models.ts add provider myprovider https://api.example.com/v1 --api-key sk-your-api-key

# 添加模型
bun ~/.pi/agent/skills/models-config/models.ts add model pox gpt-5.4 --reasoning --context-window 200000

# 测试连接
bun ~/.pi/agent/skills/models-config/models.ts test pox gpt-5.4

Test inference

bun ~/.pi/agent/skills/models-config/models.ts test pox gpt-5.4 --reasoning --thinking high

Test streaming responses

bun ~/.pi/agent/skills/models-config/models.ts test pox gpt-5.4 --stream

Batch test

bun ~/.pi/agent/skills/models-config/models.ts batch pox

Update pricing

bun ~/.pi/agent/skills/models-config/models.ts update

Validate configuration

bun ~/.pi/agent/skills/models-config/models.ts validate

## 命令详解

### list - 列出模型

```bash
bun models.ts list                    # 列出所有
bun models.ts list --provider pox    # 查看 pox 详情
bun models.ts list --json            # JSON 格式输出

add provider - Add Provider

bun models.ts add provider <name> <baseUrl> [options]

# 选项:
#   --api-key <key>     API 密钥
#   --api <type>        API 类型 (默认: openai-responses)
#   --auth-header       使用 Authorization header

# 示例:
bun models.ts add provider myprovider https://api.example.com/v1 \
  --api-key sk-your-api-key \
  --api openai-responses \
  --auth-header

add model - Add model

bun models.ts add model <provider> <modelId> [options]

# 选项:
#   --name <name>           显示名称
#   --reasoning            支持推理功能
#   --context-window <n>   上下文窗口 (默认: 128000)
#   --max-tokens <n>       最大输出 (默认: 16384)
#   --cost-input <n>       输入价格 $/M
#   --cost-output <n>      输出价格 $/M

# 示例:
bun models.ts add model pox gpt-5.4 \
  --name "GPT-5.4" \
  --reasoning \
  --context-window 200000 \
  --max-tokens 64000

test - Test connection ⭐

Supports all mainstream API types and advanced features:

bun models.ts test <provider> [modelId] [options]

# 基本测试
bun models.ts test pox gpt-5.4

# 指定测试消息
bun models.ts test pox gpt-5.4 --message "What is 2+2?"

# 测试流式响应
bun models.ts test pox gpt-5.4 --stream

# 测试推理能力
bun models.ts test pox gpt-5.4 --reasoning --thinking high

# 推理级别: minimal / low / medium / high / xhigh

Supported API types:

API type	Description	Endpoint
openai-responses	OpenAI Responses API	/responses
openai-completions	OpenAI Chat Completions	/chat/completions
anthropic-messages	Anthropic Messages API	/v1/messages
google-generative-ai	Google Gemini API	/v1beta/models/:generateContent
mistral-conversations	Mistral API	/v1/chat/completions
azure-openai-responses	Azure OpenAI	/openai/deployments/.../responses

batch - Batch test

# 批量测试所有模型
bun models.ts batch pox

# 带推理的批量测试
bun models.ts batch pox --reasoning

# 带流式的批量测试
bun models.ts batch pox --stream

rm - Delete

# 删除整个 provider
bun models.ts rm pox

# 删除指定模型
bun models.ts rm pox gpt-5.4

update - Update pricing

From models.dev get the latest pricing data:

bun models.ts update
bun models.ts update --provider pox  # 只更新 pox

template - Configuration template

# 查看 OpenAI 兼容配置模板
bun models.ts template openai-compatible

# 查看 Anthropic 兼容模板
bun models.ts template anthropic-compatible

validate - Validate JSON

bun models.ts validate

Complete configuration example

{
  "providers": {
    "pox": {
      "baseUrl": "https://api.example.com/v1",
      "apiKey": "sk-your-api-key",
      "api": "openai-responses",
      "authHeader": true,
      "models": [
        {
          "id": "gpt-5.4",
          "name": "GPT-5.4",
          "reasoning": true,
          "input": ["text", "image"],
          "output": ["text"],
          "cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 },
          "contextWindow": 200000,
          "maxTokens": 64000
        }
      ]
    }
  }
}

Configuration file location

~/.pi/agent/models.json