RAG & SEMANTIC SEARCH

Turn private knowledge into precise, grounded answers.

Use one platform for vectorizing knowledge, retrieving relevant context and generating grounded answers—with measured usage across the full pipeline.

Start building ↗Build a complete RAG pipeline
PRODUCTION ENDPOINTPOST /v1/embeddings
RAGGROUNDED ANSWERS
VECTORSEMANTIC SEARCH
1 KEYFULL PIPELINE
Production advantage

Move faster without
shrinking your options.

AIx keeps identity, integration and measured billing consistent while your workload uses the model best suited to the job.

01

One key for retrieval and generation

Create embeddings and call the answer model through one identity, usage ledger and prepaid balance.

02

Model-level cost control

Use efficient embedding and chat models for background indexing, then reserve premium reasoning for difficult queries.

03

Architecture that can evolve

Change embedding or answer models independently without replacing the surrounding API integration.

Recommended models

Start with a model
matched to the work.

Use these as practical starting points, then evaluate quality, latency and unit economics against your own traffic.

AVAILABLE ON AIX

BGE Large EN v1.5

General-purpose English retrieval embeddings

Model API details →AVAILABLE ON AIX

Llama 3.3 70B Turbo

Efficient grounded answer generation

Model API details →AVAILABLE ON AIX

Gemini 2.5 Pro

Very large-context document synthesis

Model API details →
Reference architecture

From request
to production.

  1. 01
    Chunk and embed knowledge

    Split source documents into useful semantic units and store their vectors with source metadata.

  2. 02
    Retrieve and rerank

    Embed the user query, fetch the closest passages and apply filters or reranking before generation.

  3. 03
    Generate with citations

    Send only the strongest context to the chat model and require answers to reference the supplied sources.

QUICKSTARTPOST /v1/embeddings
from openai import OpenAI

client = OpenAI(
    base_url="https://api.aix.theaimart.co/v1",
    api_key="$AIX_KEY",
)

vector = client.embeddings.create(
    model="BAAI/bge-large-en-v1.5",
    input="How does measured billing work?",
).data[0].embedding
Build a complete RAG pipeline →
FAQ

Questions, answered.

Can one AIx key power an entire RAG pipeline?+

Yes. Use the same key for embeddings and chat generation, with usage recorded in one AIx ledger.

Which embedding model should I start with?+

Choose a model based on language, retrieval quality, vector size and cost. BGE Large EN v1.5 is a strong general-purpose English starting point.

Does AIx store my documents?+

The API processes request content but AIx does not store prompts or model responses. Your application owns document storage and vector indexing.

Ship the first request today.

One key. One API contract. The AIx model network behind it.

Get your API key ↗Explore all models