Providers
Configure Empryo's 23 built-in LLM providers across cloud, gateways, subscriptions, local, and enterprise, plus any OpenAI-compatible custom API.
Copy & share
Loading sections…
Connect MCP or install the Empryo skillSection exports contain only that heading’s content. Markdown and text links fetch the selected content directly, without the rest of the page.
Empryo ships 23 built-in providers. Set one key and go.
Cloud
| Provider | Setup | Get a key |
|---|---|---|
| Anthropic | empryo --set-key anthropic sk-ant-... | console.anthropic.com |
| OpenAI | empryo --set-key openai sk-... | platform.openai.com |
empryo --set-key google ... | aistudio.google.com | |
| xAI | empryo --set-key xai xai-... | console.x.ai |
| Groq | empryo --set-key groq gsk_... | console.groq.com |
| DeepSeek | empryo --set-key deepseek sk-... | platform.deepseek.com |
| Mistral | empryo --set-key mistral ... | console.mistral.ai |
| NVIDIA NIM | empryo --set-key nim nvapi-... | build.nvidia.com |
| Fireworks | empryo --set-key fireworks ... | fireworks.ai |
| MiniMax | empryo --set-key minimax ... | platform.minimaxi.com |
| Ollama Cloud | /login ollama-cloud or empryo --set-key ollama-cloud ... | ollama.com/settings/keys |
Gateways (one key, many models)
| Provider | What you get | Get a key |
|---|---|---|
| LLM Gateway | One key for every major model, up to 30% off frontier | llmgateway.io |
| OpenRouter | 200+ models | openrouter.ai |
| Vercel AI Gateway | Unified gateway | vercel.com/ai-gateway |
Subscription-based
| Provider | Use | Setup |
|---|---|---|
| [GitHub Copilot](/docs/providers/copilot) | Your Copilot subscription | OAuth token from IDE |
| GitHub Models | Free with any GitHub PAT | empryo --set-key github-models ghp_... |
Local (free)
| Provider | Setup |
|---|---|
| Ollama | Auto-detected at localhost:11434. Set OLLAMA_HOST to override. |
| LM Studio | Auto-detected at localhost:1234. Set LM_STUDIO_URL to override. |
| OpenCode Go | empryo --set-key opencode-go ... |
| OpenCode Zen | empryo --set-key opencode-zen ... |
Enterprise
| Provider | Setup |
|---|---|
| Bedrock | export AWS_ACCESS_KEY_ID=... AWS_SECRET_ACCESS_KEY=... AWS_REGION=us-east-1 |
| Subscriptions | Claude, ChatGPT, Grok, Cursor, Devin, Gemini CLI, Antigravity, Kimi Code and Fire Pass, signed in natively. /login, no addon. Use at your own risk. See Subscriptions. |
| Proxy | Local CLIProxyAPI relay - routes Claude, GPT, Gemini, Grok via your existing chat subscriptions. Install the addon (empryo addon install proxy), then /proxy login. Use at your own risk. See addons and Proxy. Claude and ChatGPT no longer need it: see Subscriptions. |
LLM Gateway
One key for 200+ models with built-in cost tracking — Claude, GPT, Gemini, Grok, Llama, Qwen, and more. Or DevPass: a fixed-price monthly subscription with 3× the value.
- Get a key at llmgateway.io, or just type
/logininside Empryo — the browser mints one and hands it back over a local loopback. - Or save a key manually:
empryo --set-key llmgateway sk-...- Launch Empryo, press
Ctrl+L, pick any model.
The picker's gateway header says which lane is paying — Pay-as-you-go · account credits or DevPass · subscription — so a model pick is never a billing surprise. Switching lanes happens in /keys (it's a credential, not a toggle).
Pick who serves your model (upstream routing)
The gateway resells many models from several upstream vendors at different prices and output caps — the same deepseek-v4-pro costs $0.435/M input on DeepSeek and $2.40/M on Alibaba, and its output cap ranges from 64K to 393K depending on the route. By default the id is sent bare and the gateway auto-routes. On the pay-as-you-go lane you can pin the route per model:
- Terminal — in the model picker, rows with more than one route show
⇌ N. PressCtrl+Ron one to open the route chooser:Auto — gateway decidesfirst, then every upstream cheapest-first with that route's own price and output cap. Picking a route picks the model pinned to it; the active row then readsvia DeepSeek. - Desktop — gateway rows carry a
⇌button that swaps the picker to the same route list. - Headless — pin in the id itself:
empryo --headless "fix the tests" --model llmgateway/anthropic/claude-opus-5
empryo --headless "fix the tests" --model llmgateway/deepinfra/deepseek-v4-prollmgateway/<upstream>/<model> routes through that vendor; llmgateway/<model> lets the gateway decide. Pinned or bare, it is the same catalog entry — favourites, Recents, pricing and the context window all follow the model, and the pin travels with the tab (each tab can run the same model on a different route). Anthropic-pinned Claude ids additionally unlock Empryo's native tool-deferral path. Route pinning is pay-as-you-go only: a DevPass plan's routing belongs to the gateway.
OpenRouter
200+ models behind a single key, including frontier and open-weight.
- Get a key at openrouter.ai/keys.
- Save it:
empryo --set-key openrouter sk-or-...- Launch Empryo, press
Ctrl+L, pick any model.
Anything else
Any OpenAI-compatible API works via custom providers - one block of config, no code. The key goes in the same way, under the id you gave the provider:
empryo --set-key together sk-... # `id` from your providers[] entryRun it again with a second key and both are kept — a quota 429 hands the turn to the next one. See several accounts on one provider.
Switching models
Inside Empryo, press Ctrl+L to pick a provider and model. Assign different models to different tasks with /router (cheap model for search, strong for code).
empryo --list-providers # see what's configured
empryo --list-models anthropic # see available models for one provider