Provider 速查
md2html 用环境变量自动检测可用 provider。配好对应 env 就能用。
Chat（21 家，全 OpenAI-compatible）
| Provider | Env var | Default model |
|---|---|---|
| DeepSeek | DEEPSEEK_API_KEY |
deepseek-chat |
| OpenAI | OPENAI_API_KEY |
gpt-4o-mini |
| Anthropic | ANTHROPIC_API_KEY |
claude-sonnet-4-7 |
| Gemini | GEMINI_API_KEY |
gemini-2.5-flash |
| Zhipu | ZHIPU_API_KEY |
glm-4-flash |
| Moonshot (Kimi) | MOONSHOT_API_KEY |
moonshot-v1-8k |
| Qwen (通义) | DASHSCOPE_API_KEY |
qwen-plus |
| Doubao (字节) | ARK_API_KEY |
doubao-pro-32k |
| 文心 | QIANFAN_API_KEY |
ernie-4.0-8k |
| 混元 | HUNYUAN_API_KEY |
hunyuan-pro |
| 星火 | SPARK_API_PASSWORD |
spark-3.5 |
| 零一 Yi | YI_API_KEY |
yi-large |
| 阶跃 | STEP_API_KEY |
step-1 |
| xAI Grok | XAI_API_KEY |
grok-2 |
| Mistral | MISTRAL_API_KEY |
mistral-medium |
| Cohere | COHERE_API_KEY |
command-r |
| Groq | GROQ_API_KEY |
llama-3.1-70b-versatile |
| Together | TOGETHER_API_KEY |
meta-llama/Llama-3-70b-chat-hf |
| Fireworks | FIREWORKS_API_KEY |
accounts/fireworks/models/llama-v3-70b-instruct |
| OpenRouter | OPENROUTER_API_KEY |
anthropic/claude-sonnet-4-7 |
| custom | 配置 baseUrl + envKey | — |
Image（5 家）
| Provider | Env var | Default model |
|---|---|---|
| Zhipu CogView | ZHIPU_API_KEY |
cogview-3-plus |
| Gemini | GEMINI_API_KEY |
gemini-2.5-flash-image |
| OpenAI | OPENAI_API_KEY |
gpt-image-1 |
| DashScope (通义万相) | DASHSCOPE_API_KEY |
wanxiang-v2.1 |
| Doubao (即梦) | ARK_API_KEY |
doubao-image |
Publish target（4 家）
| Target | Env vars |
|---|---|
| Cloudflare Pages | CLOUDFLARE_API_TOKEN + CLOUDFLARE_ACCOUNT_ID + CLOUDFLARE_PAGES_PROJECT |
| GitHub Pages | GITHUB_TOKEN + GITHUB_REPO (+ 可选 GITHUB_BRANCH) |
| Netlify | NETLIFY_TOKEN (+ 可选 NETLIFY_SITE_ID) |
| S3-compatible | S3_ENDPOINT + S3_BUCKET + S3_REGION + AWS_ACCESS_KEY_ID + AWS_SECRET_ACCESS_KEY |
优先级 / capability negotiation
如果同时配了多个 chat provider，md2html 按 capability negotiation 自动选择能满足当前 engine requires 的 provider。
手动指定优先：在 md2html.config.json 加：
{
"providers": { "preferred": ["deepseek", "openai", "openrouter"] }
}