AI-native quantum computing SDK with 23 MCP tools for quantum simulation, algorithms, machine learning, and business applications. Exposes quantum circuits as MCP tools enabling AI assistants to perform quantum computations.
23 tools with generally clear names and descriptions, but significant gaps in schema completeness and parameter documentation. Most tools have 1-4 parameters with type info, but many lack detailed constraints, enums, or output schema documentation. Error handling is minimal, no recovery guidance or actionable error messages. Security validation (code sandboxing) is present but not reflected in tool descriptions. Descriptions average ~120 chars (within baseline 194), but parameter descriptions are sparse or missing entirely. No tool annotations (readOnlyHint/destructiveHint) despite clear risk levels. Output schemas are undocumented for all tools.
Quantum clustering using quantum k-means algorithm.
Compare MWPM (Minimum Weight Perfect Matching) vs Union-Find decoders for surface code.
Create and measure a Bell state |Φ+⟩ = (|00⟩ + |11⟩)/√2. The simplest entanglement demonstration. Should produce approximately 50% |00⟩ and 50% |11⟩.
Generate an SVG circuit diagram from Quanta circuit code.
Surface code and Willow dynamic QEC resource estimation for fault-tolerant quantum computing.
Interpret measurement results and provide quantum state analysis.
Grover's search algorithm for unstructured database search.
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields. E.g., run_circuit returns 'measurement results' but structure is undefined; monte_carlo_price returns a price but no confidence interval or Greeks breakdown.
Parameter descriptions are sparse or missing. E.g., 'graph' in qaoa_optimize has no format spec (adjacency list? edge list? dict?); 'marked_state' in grover_search lacks example format; 'noise_model' in simulate_noise lists enum values in description instead of as a formal enum constraint.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels. run_on_ibm is marked WRITE but has no destructiveHint; most tools are READ_ONLY but lack readOnlyHint. Agents cannot infer safety without annotations.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 58 | <=2025-11-25 | v2 |
List available IBM quantum computers with specifications.
Poll job status and fetch results from IBM Quantum.
List all available quantum gates (31 total) with their descriptions and matrix representations.
Quantum Monte Carlo option pricing for European and Asian options.
Compiler optimization with metrics for circuit depth reduction and gate count minimization.
Option sensitivity Greeks (Delta, Gamma, Vega, Theta, Rho) using quantum differentiation.
QAOA combinatorial optimization for MaxCut and other problems.
Syndrome diagnosis and error lookup for quantum error correction.
Quantum ML classification using quantum feature maps and ansatz presets.
Reasoning complexity feedback for AI agents on quantum circuit design.
Execute a quantum circuit and return measurement results. Write circuit code using Quanta SDK syntax. The code should define a circuit function using the @circuit decorator and return it.
Run circuit on IBM Quantum hardware (Heron r3, 156 qubits).
Shor's factoring algorithm for integer factorization.
Run circuit with noise model simulation.
Surface code quantum error correction simulation.
Transpile circuit to Heron, Willow or IonQ native ISA.
Error handling is absent. Code validation in _validate_code() raises ValueError but no tool documents what errors are possible or how to recover. E.g., if run_circuit fails due to blocked code, the LLM gets a raw error with no guidance on what patterns are allowed.
Numeric parameters lack bounds. E.g., 'shots' defaults to 1024 but no min/max documented; 'num_qubits' in grover_search has no upper limit; 'error_rate' in simulate_noise has no range (0-1? percentage?). Unbounded params invite invalid LLM inputs.