Introduction
Alvin is an AI workspace with hundreds of models from every major provider, and routing that picks the right one for you.
Alvin gives you one place to work with AI. Chat, generate images and video, build documents and boards, and let a team of specialist agents handle the parts you'd rather not think about — across hundreds of models from OpenAI, Anthropic, Google, Meta, xAI and more.
You never have to pick a model. Alvin reads what you're asking for and routes it to the one that fits.
Why Alvin?
One workspace
Every major provider behind a single account. No juggling logins, keys, or separate bills.
Routing that thinks
Alvin reads your intent and sends each prompt to the model best suited to it — for quality and for cost.
A team of specialists
Victoria makes images and video. Chris builds and previews code. Peter runs the board. Ava wires up workflows.
Built for teams
Organizations, projects, shared memory, real-time collaboration, and usage you can actually see.
Two ways in
Alvin Studio
The workspace at studio.tryalvin.com — where you chat, create, and collaborate.
- Chat — conversation with automatic model routing, web search, and voice
- Canvases — documents, sheets, code, boards, planners, and calendars built right in the conversation
- Model Hub — browse, compare, and benchmark the model catalog
- Projects — group your work, share context, track spend
The Elyxir API
Alvin Studio is built on the Elyxir platform. The same models and routing are available programmatically through an OpenAI-compatible API at api.elyxir.ai — point your existing OpenAI SDK at it and change the key.
curl -X POST https://api.elyxir.ai/v1/chat/completions \
-H "Authorization: Bearer elyxir_your_api_key" \
-H "Content-Type: application/json" \
-d '{"model": "alvin", "messages": [{"role": "user", "content": "Hello!"}]}'model: "alvin" hands the request to Alvin's router. Pass a specific model ID instead — gpt-5.5, claude-opus-4.8, gemini-3.5-flash — to bypass routing.
Getting started
- Create an account at studio.tryalvin.com
- Start a chat — Alvin picks the model, you don't have to
- Generate an API key in Settings > API Keys when you're ready to build
Supported providers
| Provider | Models | Capabilities |
|---|---|---|
| OpenAI | GPT-5.5, GPT-5.4, GPT-4.1, GPT-4o | Chat, Vision, Embeddings, Image, Audio |
| Anthropic | Claude 4.8, 4.7, 4.6, 4.5 | Chat, Vision, Long context, Agentic coding |
| Gemini 3.5, 3.1, 3, 2.5 | Chat, Vision, Multimodal, Image, Video | |
| xAI | Grok 4.3, Grok 4 | Chat, Reasoning |
| Meta | Llama 4, Llama 3.3 | Chat, Code, Open weights |
| Mistral AI | Mistral Large 3, Small 4, Codestral | Chat, Code |
| DeepSeek | DeepSeek V4 Pro, V3.2 | Reasoning, Code |
| Cohere | Command A+, Command A | Chat, RAG, Embeddings |
| Perplexity | Sonar Deep Research, Sonar | Chat, Web search |
| Fireworks AI | Qwen 3.5, Kimi K2.6, DeepSeek V4 | Fast inference |
| ElevenLabs | eleven-v3, Scribe v2 | Text-to-speech, Transcription |
See Supported models for the full catalog with pricing.
Next steps
- Quickstart — get started in 5 minutes
- Studio overview — explore the workspace
- API reference — build on the Elyxir platform