API Overview

The SolidRusT AI API provides OpenAI-compatible endpoints for AI inference.

Base URL

https://api.solidrust.ai/v1

Available Endpoints

Chat & Inference

Endpoint	Method	Description
/v1/chat/completions	POST	Generate chat completions
/v1/embeddings	POST	Create text embeddings
/v1/models	GET	List available models

Agent (Tool-Augmented Chat)

Endpoint	Method	Description
/v1/agent/chat	POST	Chat with built-in RAG tools + custom tools
/v1/agent/tools	GET	List available tool schemas
/v1/agent/health	GET	Agent health check

Data Layer (RAG)

Endpoint	Method	Description
/data/v1/query/semantic	POST	Semantic vector search
/data/v1/query/hybrid	POST	Combined semantic + knowledge graph
/data/v1/query/knowledge-graph	POST	Entity relationship queries
/data/v1/ingest/document	POST	Ingest single document
/data/v1/ingest/batch	POST	Batch document ingestion

Request Format

All requests must include:

• Content-Type: application/json header
• Authorization: Bearer YOUR_API_KEY header
• JSON request body (for POST requests)

Example Request

curl https://api.solidrust.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "vllm-primary",
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

Response Format

All responses are JSON objects. Successful responses include the requested data. Error responses follow a standard format:

{
  "error": {
    "message": "Description of the error",
    "type": "error_type",
    "code": "error_code"
  }
}

OpenAPI Specification

Interactive API documentation is available:

• Swagger UI: api.solidrust.ai/data/docs
• ReDoc: api.solidrust.ai/data/redoc
• OpenAPI JSON: api.solidrust.ai/data/openapi.json

These endpoints document the Data Layer API (RAG, agent, ingestion). The chat completions and embeddings endpoints follow the standard OpenAI API specification.

Rate Limiting

See Rate Limits for information about request limits.