Provider reference
Endpoints, key pages, sign-up notes and starter models for all eight providers.
The eight providers rdn-swarm supports: where to get a key, what the app connects to, and what to watch out for when signing up. Every provider works for every role, and a run can mix them freely.
At a glance#
- In the app
- Anthropic
- API host
api.anthropic.com- Key page
- console.anthropic.com
- In the app
- OpenAI
- API host
api.openai.com- Key page
- platform.openai.com
- In the app
- Gemini
- API host
generativelanguage.googleapis.com- Key page
- aistudio.google.com
- In the app
- DeepSeek
- API host
api.deepseek.com- Key page
- platform.deepseek.com
- In the app
- xAI
- API host
api.x.ai- Key page
- console.x.ai
- In the app
- Meta
- API host
api.meta.ai- Key page
- ai.developer.meta.com
- In the app
- Kimi
- API host
api.moonshot.ai- Key page
- platform.kimi.ai
- In the app
- Qwen
- API host
dashscope-intl.aliyuncs.com- Key page
- home.qwencloud.com
The API host is shown on each provider's page in Settings. It's the only address that provider's key is ever sent to.
Anthropic#
Sign in to the Anthropic Console, add billing or credits, then generate a key under Settings → API Keys. A Claude.ai subscription doesn't include API access. Starter model: claude-haiku-4-5, which also reads images, so it can be used for Visual QA.
OpenAI#
Sign in on platform.openai.com. This is separate from ChatGPT, and a ChatGPT subscription gives no API access. Verify your email and phone, add a billing method, then create a key on the API keys page. Starter model: gpt-5-mini.
Google Gemini#
Open Google AI Studio and create a key with your Google account. A free tier exists; the paid tier needs a linked Google Cloud billing project. Free-tier rate limits are lower, and several agents calling at once reach them sooner. Starter model: gemini-2.5-flash-lite.
DeepSeek#
Sign in to the DeepSeek platform, top up your balance, then create a key. DeepSeek offers an off-peak discount at set UTC hours, and rdn-swarm prices each call by the time it started. Starter model: deepseek-flash.
xAI#
Sign in to the xAI Console with your X or email account. An X Premium subscription gives no API access. Add credits under Billing, then create a key on the API Keys page. Starter model: grok-4.3.
Meta#
Sign in to the Meta Model API portal with a Meta developer account and create a key under API keys. The hosted models are the Muse family; Llama is a separate open-weights download and isn't offered here. Starter model: muse-spark-1.3.
Kimi#
Sign in to the international Kimi Open Platform, top up your balance, then create a key under API Keys. A key from the China platform (.cn) is rejected by the international endpoint rdn-swarm uses. Starter model: kimi-k2.7-code.
Qwen#
Sign in at home.qwencloud.com and create a key on the API Keys page. rdn-swarm uses the international (Singapore) endpoint. Keys are region-specific, and a key minted for another region is rejected with an incorrect API key error rather than a clear region error. Starter model: qwen3.8-flash.
Prices#
rdn-swarm's model list and prices come from Rdn Labs' public price feed, built from each provider's own pricing pages. The current files are at swarm.reidell.net/prices/<provider>.json. Only models with a published price appear in the app. See Providers, models and pricing.