Agentic Development¶

TensorCircuit-NG is the world’s first AI-native quantum programming platform, purpose-built for agentic research and automated scientific discovery.

🚀 Experience Agent-Native Discovery

Click here to see how AI agents autonomously solve complex quantum problems in TensorCircuit-NG.

agent_landing/index.html

Why Work Within the Repository?¶

To write scripts and applications efficiently with AI coding agents (e.g., ClaudeCode, Cursor, Codex, Antigravity, Gemini-CLI, OpenCode), we strongly recommend working directly within the local repository.

Rich Context: The repository contains over 100 scripts in examples/ and extensive test cases in tests/. These provide essential references that significantly reduce AI hallucinations and help the agent understand idiomatic usage.
Built-in Rules: We provide a dedicated AGENTS.md file. It serves as the “handbook” (similar to CLAUDE.md) for AI agents, defining coding standards and best practices.
Specialized Agentic Skills: The .agents/skills/ directory contains specialized workflows to guide AI assistants on complex, multi-step tasks.

Specialized Agentic Skills¶

TensorCircuit-NG includes built-in agentic skills that can be activated by compatible AI agents to perform advanced tasks:

arxiv-reproduce: Autonomously reproduces arXiv papers with standardized output and code quality validation.
performance-optimize: Scientific execution and memory optimization workflow (JAX scanning, vectorized parallelism, etc.).
tc-rosetta: End-to-end framework translation (from Qiskit, PennyLane, etc.) with intrinsic mathematical intent rewriting.
tutorial-crafter: Transforms raw scripts into comprehensive, narrative-driven educational tutorials.
demo-generator: Transforms scripts into interactive, high-performance Streamlit GUI applications.
code-reviewer: Autonomously reviews and refactors code for mathematical correctness and performance.
meta-explorer: High-intensity autonomous research agent for circuit architecture and optimization strategy discovery (VQE, QML, QAOA, etc.).

Recommended Workflow¶

Clone the repository:

git clone https://github.com/tensorcircuit/tensorcircuit-ng.git

Switch to a local playground branch:
```
git checkout -b my-playground
```
Open the repository folder in your AI IDE: Start writing TC-NG-based scripts using natural language instructions.

By integrating extreme performance with an autonomous, intent-driven AI workflow, TensorCircuit-NG empowers researchers to transition from manual coding to automated scientific discovery.

AI-Native Documentation¶

One can also refer to AI-native docs for tensorcircuit-ng: Devin Deepwiki, Google Code Wiki, and Context7 MCP.