multi-provider-setup

One CLI to configure every AI coding agent with 165 LLM providers.

Add API keys, health-check providers, and auto-generate configs for OpenCode, Codex CLI, Claude Code, and Antigravity CLI — all from one interactive terminal session.

npx multi-provider-setup
# or
bunx multi-provider-setup

Features

• 165 providers — Groq, OpenAI, Anthropic, Together, Fireworks, DeepInfra, Cerebras, Replicate, Ollama, LM Studio, and 155+ more
• 47 no-credit-card providers — free tiers that don't require billing info
• 99 permanent free tiers — providers with always-free access
• 18 local engines — Ollama, llama.cpp, vLLM, LocalAI, LM Studio, and more (zero cost)
• Zero-config health check — pings each provider's /v1/models or /v1/messages, measures latency, and discovers available models
• 4 target agents — generate configs for OpenCode, Codex CLI, Claude Code, and Antigravity CLI simultaneously
• Interactive TUI — searchable provider list, API key prompts, live health check spinner, config preview
• Batteries included — works immediately with bun, no additional runtime deps

Usage

Run the CLI:

npx multi-provider-setup

Or if you have bun installed:

bunx multi-provider-setup

Or run globally:

npm install -g multi-provider-setup
mps

The 5-Step Workflow

Step 1 ── Select providers (searchable checklist)
     │       Pick from 165 providers — type to filter
     ▼
Step 2 ── Enter API keys
     │       Prompts only for providers that need keys
     ▼           (local engines are auto-skipped)
Step 3 ── Health check
     │       Pings each provider, measures latency,
     ▼           discovers available models
Step 4 ── Select target agents
     │       OpenCode, Codex CLI, Claude Code,
     ▼           Antigravity CLI — or all at once
Step 5 ── Configs generated
             Files written: opencode.json,
             ~/.codex/config.toml, ~/.claude/settings.json,
             ~/.config/antigravity/config.toml

Output Files

| Agent | File | Description | |-------|------|-------------| | OpenCode | ./opencode.json | Provider config with model lists | | Codex CLI | ~/.codex/config.toml | [model_providers] entries | | Claude Code | ~/.claude/settings.json + ~/.claude/env.sh | Settings JSON + env var helpers | | Antigravity CLI | ~/.config/antigravity/config.toml | Provider entries |

Supported Providers (165)

Cloud Providers — Free Tier (47, no CC needed)

| Provider | baseURL | Free Tier | |----------|---------|-----------| | Groq | api.groq.com/openai/v1 | 30 RPM, 14,400 RPD | | Cerebras | api.cerebras.ai/v1 | 1M tokens/day | | Google Gemini | generativelanguage.googleapis.com/v1beta | 1,500 RPD (Flash) | | GitHub Models | models.github.ai/inference | Free for GitHub accounts | | Mistral AI | api.mistral.ai/v1 | 1B tokens/month | | Cloudflare Workers AI | api.cloudflare.com/client/v4/accounts/{account}/ai/v1 | 10K neurons/day | | Hugging Face | router.huggingface.co/v1 | $0.10/mo credits | | Cohere | api.cohere.com/v2 | 1K requests/mo trial | | AI21 Labs | api.ai21.com/studio/v1 | 200 RPM, 10 RPS | | OpenRouter | openrouter.ai/api/v1 | 20+ free models | | DeepSeek | api.deepseek.com/v1 | 10M free tokens | | Zhipu AI / GLM | open.bigmodel.cn/api/paas/v4 | Free Flash models | | SiliconFlow | api.siliconflow.cn/v1 | 1K RPM on free models | | ... and 34 more | | |

Cloud Providers — Paid / Trial Credits

OpenAI, Anthropic, xAI/Grok, Together AI, Fireworks AI, Baseten, Nebius AI, DeepInfra, Novita AI, Perplexity, Replicate, Modal, SambaNova, Scaleway, Hyperbolic, Inference.net, Upstage, and 40+ more.

Local & Self-Hosted (18)

Ollama, LM Studio, llama.cpp, vLLM, LocalAI, SGLang, Text Generation WebUI, TabbyAPI, KoboldCpp, Aphrodite Engine, GPT4All, Jan.ai, Tabby AI, MindSpore, ONNX Runtime, AegisGate, FreeLLMAPI, Atomic Chat.

Chinese & Asian Providers

ModelScope, DashScope / Alibaba, MiniMax, Moonshot / Kimi, Baichuan, Stepfun, Tencent Hunyuan, ByteDance Volcengine / Ark, SenseTime, Yi / 01.AI, Gitee AI, Xiaomi Mimo, and more.

European & GDPR-Compliant

OVHcloud, Scaleway, Nebius, Aleph Alpha, AiQu, Berget AI, Cortecs AI, EUrouter, Infercom, Verda.

Gateways & Aggregators

OpenRouter, Portkey, Helicone, LiteLLM, Cloudflare AI Gateway, Vercel AI Gateway, Routeway, Adaline, LLM Gateway, TokenRouter, FastRouter, ZenMux.

Development

# Clone
git clone https://github.com/YOUR_USER/multi-provider-setup
cd multi-provider-setup

# Install dependencies
bun install

# Type check
bun run typecheck       # tsc --noEmit

# Build
bun run build           # bun build ./src/index.ts --outdir ./dist --target node

# Run locally
bun run dev             # bun run ./src/index.ts

Project Structure

multi-provider-setup/
├── src/
│   ├── index.ts              # CLI entry point (5-step workflow)
│   ├── providers.ts          # 165 provider definitions
│   ├── health.ts             # Health check logic
│   └── configs/
│       ├── index.ts          # Re-exports + generateAllConfigs
│       ├── opencode.ts       # opencode.json generator
│       ├── codex.ts          # ~/.codex/config.toml generator
│       ├── claude.ts         # Claude Code settings + env vars
│       └── antigravity.ts    # Antigravity CLI config generator
├── scripts/
│   ├── sync-modelscope-models.py
│   └── opencode-with-modelscope
├── dist/
│   └── index.js              # Built output (~115KB)
├── package.json
├── AGENTS.md
└── README.md

License

MIT