- API Type
- REST API with OpenAPI specification. Endpoints for completion, chat completion, JSON completion, embeddings, tokenization, detokenization, evaluation, and explanation.
- Authentication
- API token (AA_TOKEN or ALEPHALPHA_API_KEY). Passed via Client initialization in SDKs or Authorization header.
- Webhooks
- No webhook support mentioned in public documentation.
- SDKs
- Official SDKs: Python (aleph-alpha-client), Rust (aleph_alpha_api), JavaScript/TypeScript (intelligence-layer-sdk). Community integrations via LiteLLM, LlamaIndex, Cognigy.
- Documentation
- Comprehensive API documentation at docs.aleph-alpha.com with OpenAPI specs, code examples, and model references. PhariaInference and PhariaAI-specific guides available.
- Sandbox
- No public sandbox mentioned. Testing requires valid API token with usage limits based on subscription.
- SLA
- Not publicly documented. Enterprise customers likely receive dedicated SLAs via sales contact.
- Rate Limits
- Token-based usage limits per subscription tier. Specific RPM/TPM limits not publicly documented.
- Use Cases
- Text completion, chat/multimodal conversations, semantic embeddings, JSON generation, RAG/QA pipelines, evaluation pipelines, enterprise AI workflows via PhariaAI.