Provider-Based Model System
Nadoo AI uses a capability-based provider system that lets you connect to 12+ AI model providers through a unified interface. Each provider exposes one or more capabilities — such as LLM chat, embeddings, speech-to-text, or image generation — and your applications automatically route requests to the right model based on what they need.You can configure multiple providers simultaneously. Nadoo AI will use the provider assigned to each application, giving you full control over cost, latency, and capability tradeoffs.
Supported Providers
| Provider | Models | Capabilities |
|---|---|---|
| OpenAI | GPT-4o, GPT-4, GPT-3.5-turbo | LLM, Embedding, TTS, TTI, Vision |
| Anthropic | Claude 3.5 Sonnet, Claude 3 Opus, Claude 3 Haiku | LLM |
| Azure OpenAI | GPT-4, GPT-3.5 deployments | LLM, Embedding |
| AWS Bedrock | Claude, Llama, Titan | LLM, Embedding |
| Google Gemini | Gemini Pro, Gemini Flash | LLM, Vision |
| Google Vertex AI | PaLM, Gemini | LLM, Embedding |
| Ollama | Llama, Mistral, CodeLlama | LLM, Embedding |
| vLLM | Any open-source model | LLM |
| OpenRouter | 100+ models | LLM |
| OpenAI-Compatible | Any OpenAI-compatible API | LLM |
Provider Capabilities
Each provider can support one or more of the following capabilities:| Capability | Description |
|---|---|
| LLM | Chat completion and text generation with streaming support |
| Embedding | Convert text into vector representations for semantic search |
| STT | Speech-to-text transcription from audio input |
| TTS | Text-to-speech audio generation |
| TTI | Text-to-image generation (e.g., DALL-E) |
| Image Understanding | Vision capabilities for analyzing images |
| Reranking | Re-score and reorder search results for relevance |
How to Configure a Provider
Add Your Provider
Click the provider you want to configure and enter your API key. Some providers require additional fields such as an endpoint URL or deployment name.
Provider Quick Links
OpenAI
GPT-4o, embeddings, DALL-E, Whisper, and TTS. The most feature-rich provider with support for all capability types.
Anthropic
Claude 3.5 Sonnet, Opus, and Haiku. Best-in-class for complex reasoning, long documents, and code generation.
Choosing the Right Provider
Best for general-purpose chat
Best for general-purpose chat
OpenAI GPT-4o or Anthropic Claude 3.5 Sonnet — both deliver strong performance across a wide range of tasks. GPT-4o offers lower latency; Claude 3.5 Sonnet excels at nuanced reasoning.
Best for cost-sensitive workloads
Best for cost-sensitive workloads
GPT-4o-mini, Claude 3 Haiku, or Ollama (self-hosted). These models offer excellent quality-to-cost ratios for high-volume applications.
Best for private / air-gapped deployments
Best for private / air-gapped deployments
Ollama or vLLM — run open-source models entirely on your own infrastructure with no data leaving your network.
Best for embeddings and RAG
Best for embeddings and RAG
OpenAI text-embedding-3-small/large or Azure OpenAI embeddings. For self-hosted options, Ollama supports embedding models as well.
Provider availability may vary depending on your deployment type. Self-hosted deployments can connect to any provider with network access. Cloud-hosted deployments on Nadoo Cloud come with OpenAI and Anthropic pre-configured.