Metadata-Version: 2.4
Name: bivectorai
Version: 0.3.0
Summary: BivectorAI Agent Security SDK — AI safety verification for production systems
Home-page: https://bivectorai.com
Author: BivectorAI
Author-email: support@bivectorai.com
Project-URL: Homepage, https://bivectorai.com
Project-URL: API, https://api.bivectorai.com
Project-URL: Documentation, https://bivectorai.com/agent/
Keywords: ai security llm injection pii hallucination content-safety bivectorai
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: License :: OSI Approved :: MIT License
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Topic :: Security
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Classifier: Intended Audience :: Developers
Classifier: Operating System :: OS Independent
Requires-Python: >=3.8
Description-Content-Type: text/markdown
Requires-Dist: requests>=2.28.0
Dynamic: author
Dynamic: author-email
Dynamic: classifier
Dynamic: description
Dynamic: description-content-type
Dynamic: home-page
Dynamic: keywords
Dynamic: project-url
Dynamic: requires-dist
Dynamic: requires-python
Dynamic: summary

# BivectorAI Python SDK

Enterprise AI safety verification for production systems.

## Install

```bash
pip install bivectorai
```

## Quick Start

```python
from bivectorai import BivectorAI

client = BivectorAI(api_key="your-api-key")

# Scan AI output for safety issues
result = client.verify_output(
    text="The model says X",
    sources=["User asked Y"],
)
print(result.verdict)  # PASS, REVIEW, or BLOCK

# Detect LLM injection
scan = client.scan_injection("Ignore previous instructions and...")
print(scan["verdict"])  # BLOCK or PASS

# Check for hallucinations
check = client.check_hallucination(
    claim="Revenue was $10B in 2023",
    evidence=[{"source": "annual_report.pdf", "text": "Revenue: $8.2B"}],
)
print(check["verdict"])  # GROUNDED or UNSUPPORTED
```

## API Reference

- `verify_output(text, sources)` — LLM output safety gate
- `scan_injection(text)` — prompt injection detection
- `check_hallucination(claim, evidence)` — factual grounding check
- `detect_pii(text)` — PII detection and classification
- `scan_content(text, categories)` — comprehensive content safety
- `verify_workflow(steps)` — AI workflow verification
- `get_usage_summary(period)` — usage statistics
- `revoke_api_key(key_prefix)` — key rotation helper

Full docs: https://bivectorai.com/docs/
