French version below
Enterprise-ready Generative AI API Gateway | Open Source | Sovereign Infrastructure
Developed by the French Government 🇫🇷
Albert API is an enterprise-ready open-source gateway for deploying generative AI models on your infrastructure:
- 🚦 Robust API Gateway: Load balancing, authentication, and seamless integration with OpenAI, vLLM, HuggingFace TEI.
- 📚 Advanced Features: Built-in Retrieval-Augmented Generation (RAG), OCR, audio transcription, and more.
- 🌐 Open Standards: Compatible with OpenAI APIs, LangChain, and LlamaIndex.
- 🛠️ Deployment Flexibility: Host generative AI securely on your own infrastructure, ensuring full data sovereignty.
-
Unified Access: Single API gateway for multiple generative AI model backends:
- OpenAI (Language, Embeddings, Reranking, Transcription)
- vLLM (Language)
- HuggingFace TEI (Embeddings, Reranking)
- RAG Integration: Efficiently query vector databases using Elasticsearch or Qdrant.
- Audio & Vision: Transcribe audio (Whisper) and perform OCR on PDF documents.
- Enhanced Security: Built-in API key authentication.
| Feature | Albert API ✅ | LiteLLM | OpenRouter | OpenAI API |
|---|---|---|---|---|
| Fully Open Source | ✔️ | Partially | ❌ | ❌ |
| Data Sovereignty | ✔️ | ✔️ | ❌ | ❌ |
| Multiple AI Backends | ✔️ | ✔️ | ✔️ | ❌ |
| Built-in RAG | ✔️ | ❌ | ❌ | ❌ |
| Built-in OCR | ✔️ | ❌ | ❌ | ❌ |
| Audio Transcription | ✔️ | ❌ | ❌ | ✔️ |
| Flexible Deployment | ✔️ | ✔️ | ❌ | ❌ |
| OpenAI Compatibility | ✔️ | ✔️ | ✔️ | ✔️ |
Deploy Albert API quickly on your own infrastructure:
Explore practical use cases:
- Chat Completions
- Multi-Model Access
- Retrieval-Augmented Generation (RAG)
- Knowledge Database Import
- Audio Transcriptions
- PDF OCR
Albert API thrives on open-source contributions. Join our community!
API open source pour modèles d'IA générative | Infrastructure souveraine
Albert API, porté par l'OPI de la DINUM, est le service d'IA générative de référence de l'État français, homologué pour des traitements sécurisés. Il propose une solution prête pour la production destinée à l’hébergement souverain et performant d’IA génératives avancées sur votre infrastructure.
- 🔐 Sécurité et souveraineté des données
- 🧩 API unique compatible OpenAI, vLLM et HuggingFace
- 🔎 Recherche avancée par RAG et vector stores
Consultez la documentation ou déployez rapidement votre instance via le guide de déploiement.
