TurboAgents Demo

A small standalone demo repo for trying TurboAgents outside the main source tree.

What this repo shows

• TurboQuant-style compressed retrieval and reranking through TurboAgents
• standalone TurboAgents with LanceDB
• standalone TurboAgents with SurrealDB
• optional SuperOptiX framework integrations using DSPy, Pydantic AI, and OpenAI Agents SDK

In this repo, the TurboQuant part is the compressed vector path configured through the TurboAgents adapters. The demos use low-bit compressed indexes and rerank the candidate set before returning the final hits.

Why the demo scripts are committed

The standalone demo code is checked into this repo on purpose.

That is better than generating it on demand because:

• users can read the exact code before they run anything
• the demo stays stable across recordings and launch posts
• the repo works as a public reference implementation, not just a generator
• the runner script can stay focused on execution and summary output

The checked-in superoptix-demo-workspace/ directory is there for inspection. It gives people a concrete reference for generated SuperOptiX code.

Runtime execution does not reuse that folder. The demo runner creates a separate superoptix-demo-runtime/ workspace so local runs do not overwrite the reference output.

Prerequisites

Before running this demo, make sure these local services are up:

• Ollama is running
• Docker is running
• a SurrealDB container is available on port 8000
• the Ollama model qwen3.5:9b is already pulled if you want the SuperOptiX integrations

Useful checks:

ollama list
curl http://127.0.0.1:11434/api/tags
docker ps

Quick start

git clone https://github.com/SuperagenticAI/turboagents-demo.git
cd turboagents-demo
uv sync
uv run python scripts/run_demo.py

That runs the standalone LanceDB and SurrealDB demos, shows the TurboQuant-style compressed retrieval path in action, and writes a summary to results/summary.md.

Optional SuperOptiX integrations

For the full integration path, install the extra and run:

uv sync --extra superoptix
uv run python scripts/run_demo.py --with-superoptix

This will bootstrap a local SuperOptiX workspace inside this repo, pull the packaged demo agents, compile them, and run:

• DSPy
• Pydantic AI
• OpenAI Agents SDK

That runtime workspace is created under superoptix-demo-runtime/. The checked-in superoptix-demo-workspace/ folder stays untouched so people can inspect the generated code separately.

Links

• TurboAgents docs: https://superagenticai.github.io/turboagents/
• TurboAgents repo: https://github.com/SuperagenticAI/turboagents
• TurboAgents website: https://super-agentic.ai/turboagents
• SuperOptiX integration page: https://superoptix.ai/turboagents
• SuperOptiX repo: https://github.com/SuperagenticAI/superoptix