Helps us route you to the right plan and pre-bake your savings model.
Tokani works with OpenAI, Anthropic, Azure OpenAI, AWS Bedrock, Google Vertex (Gemini), Groq, Together, Fireworks, DeepSeek, Mistral, Cerebras, xAI Grok, Perplexity Sonar, OpenRouter, AI21 Jamba, and any OpenAI-compatible self-hosted endpoint (vLLM / Ollama / TGI / NVIDIA NIM). Not currently supported: Replicate, native Hugging Face Inference API, IBM watsonx, and proprietary in-house endpoints without an OpenAI-compatible surface — let us know in the notes if that's your stack.