Use the model you want
Connect with an API key, a subscription you already have, a gateway such as OpenRouter, or models on your own machine. You pay your provider directly.
Every provider, and how it connects
9 of 9 providers
- LLM GatewaySign inPay-as-you-go credits, or the DevPass monthly planLLM Gateway (llmgateway.io) puts many model providers behind one account and one API, with cost tracking and failover. Empryo has LLM Gateway login built in: run
/login, approve in your browser, and the key is saved for you, with no copying. You can pay as you go with credits, or use DevPass, a fixed-price monthly plan. Once you are signed in, you can use any of its models in any task-router slot.claude-sonnet-4-6claude-haiku-4-5gpt-5.5gemini-3.1-proSetup guide - Anthropic (Claude)API keyAnthropic's own API prices, on one billAnthropic makes the Claude models. Anthropic is built into Empryo: save your API key once with
--set-keyand every Claude model your key can access is available for chat, code and every task-router slot.claude-opus-4-8claude-opus-4-7claude-opus-4-6claude-sonnet-4-6Setup guide - OpenAIAPI keyOpenAI's own API pricesOpenAI is built into Empryo. Save one API key and you can use OpenAI's GPT models for chat, code and any task-router slot. You can also set how much reasoning each model does.gpt-5.5gpt-5.4gpt-5.4-minigpt-5.4-nanoSetup guide
- Google (Gemini)API keyGemini API prices, with low-cost Flash modelsGoogle Gemini is built into Empryo. Save one API key and you can use any Gemini model your account can access, from the Pro models to the cheaper, faster Flash models, in any task-router slot. Flash models are a good fit for high-volume jobs like context compaction and read-only research.Gemini 3.1 Pro (gemini-3.1-pro-preview)Gemini 3.1 Flash Lite (gemini-3.1-flash-lite)Gemini 3 Flash (gemini-3-flash-preview)Gemini 2.5 Pro (gemini-2.5-pro)Setup guide
- OpenRouterAPI keyOne balance for models from many labsOpenRouter is built into Empryo. One OpenRouter key gives you models from many labs behind a single OpenAI-compatible API. Save the key once and switch between Claude, GPT, Gemini, Grok and open-weight models inside Empryo without separate accounts.anthropic/claude-opus-4.5anthropic/claude-sonnet-4.5anthropic/claude-haiku-4.5openai/gpt-5.1Setup guide
- GroqAPI keyGroq's own API pricesGroq runs open-weight models on its own LPU chips, which are built for fast responses. Groq is built into Empryo. Use it for the parts of the agent's work where speed matters most, such as search, summaries and cleanup. Save one key and all of Groq's models are available.openai/gpt-oss-120bopenai/gpt-oss-20bllama-3.3-70b-versatilellama-3.1-8b-instantSetup guide
- Ollama (local models)On your machineNothing per token: it runs on your hardwareRun open-weight models on your own machine and use them in Empryo. Ollama is built into Empryo and is detected at
localhost:11434, with no API key and nothing sent to the cloud. With a local model, your whole session can run offline with no per-token cost.qwen3-coder (30b, 480b), tool-calling, long-context codingqwen3 (0.6b, 235b), general, with tools + thinking variantsgpt-oss (20b, 120b), OpenAI open-weight, reasoning + toolsdevstral (24b), tuned for coding agentsSetup guide - GitHub CopilotSign inThe Copilot plan you already pay for, with no extra per-token billIf you already pay for GitHub Copilot, Empryo can use the models in your plan, such as Claude, GPT and Gemini, with no separate API bill. Sign in with your GitHub account from Empryo. This integration is not official: it uses the same API as the Copilot editor extensions, which GitHub can change at any time.Claude Opus 4.8Claude Opus 4.6Claude Sonnet 4.6Claude Sonnet 4.5Setup guide
- Custom / OpenAI-compatibleAny compatible APIWhatever your endpoint charges, or nothing for a local serverAny API that uses the OpenAI
/v1/chat/completionsformat can become a provider in Empryo with one block of JSON config, with no plugin or code. Use it for Together, Fireworks, Cerebras, DashScope, a company gateway, or a local vLLM or llama.cpp server. Once added, it works like a built-in provider, including in the task router.meta-llama/Llama-3.3-70B-Instruct (Together, Fireworks)deepseek-ai/DeepSeek-V3deepseek-ai/DeepSeek-R1 (reasoning)Qwen/Qwen3-235B-A22B (DashScope / OpenAI-compatible mode)Setup guide
Four ways to connect
API key
Paste a key once. Empryo keeps it in your system keychain, and you pay the provider for what you use.
5 providers
Sign in
Approve Empryo in your browser and use a plan or credits you already have. There is no key to copy.
2 providers
On your machine
Run open models on your own computer. No key, no per-token bill, and nothing has to leave your machine.
1 provider
Any compatible API
One block of JSON config turns any OpenAI-compatible endpoint into a provider, including your own server.
1 provider
Install, then pick a model
Install Empryo, connect one provider, and press Ctrl+L to choose a model.
curl -fsSL empryo.com/install.sh | bash