ainfera-llamaindex — LlamaIndex + Ainfera Routing

LlamaIndex integration + Ainfera Routing. RAG pipelines, agents, and workflows — inference routed through Ainfera with signed audit per turn.

GTM publish-only companion repo (minimal SDK sample + curl). Production LlamaIndex apps point OpenAI-compatible settings at Ainfera the same way.

Quickstart (curl only)

git clone https://github.com/ainfera-ai/ainfera-llamaindex
cd ainfera-llamaindex
./curl-example.sh   # signup → inference → audit verify

Quickstart (OpenAI SDK — same transport LlamaIndex uses)

pip install -r requirements.txt
export AINFERA_API_KEY=ai_infera_...  # https://app.ainfera.ai/signup
python main.py

LlamaIndex + Ainfera

Point LlamaIndex’s OpenAI-compatible LLM at Ainfera (exact class names vary by LlamaIndex release). The reliable pattern is environment variables consumed by LlamaIndex and the upstream OpenAI client:

export OPENAI_API_KEY="$AINFERA_API_KEY"
export OPENAI_BASE_URL="https://api.ainfera.ai/v1"

Alternatively set api_key and base URL (api_base / azure_endpoint equivalents depend on version) on your chosen OpenAI LLM wrapper from llama_index.* — mirror python main.py, which uses the official OpenAI SDK with base_url="https://api.ainfera.ai/v1".

What this shows

• One Agent Card across providers (L1)
• Routed inference with per-call budget caps (L3)
• Hash-chained receipts per LLM turn (L4)

Other frameworks

• ainfera-letta — Letta + memory blocks
• ainfera-langgraph — LangGraph
• ainfera-mcp — Claude Desktop + Cursor
• ainfera-openai-compatible — universal wedge

Apache 2.0. © Ainfera Inc. 2026.