2.0.0

aicompliant-discover v2.0.0

Enterprise-grade AI system auto-discovery SDK for Python.

v2.0.0 uses explicit wrapping (no monkey-patching, no global state mutation).

Installation

pip install aicompliant-discover

Quick Start

from aicompliant_discover import AIDiscovery
import anthropic

discovery = AIDiscovery(
    api_key="dk_your_api_key",
    endpoint="https://aicompliant.ai/v1/discovery/llm-usage",
)

# Wrap your Anthropic client — automatic usage tracking
client = discovery.wrap_anthropic(anthropic.Anthropic())

msg = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello"}],
)
# Usage automatically reported to AICompliant!

Supported Providers

Anthropic

client = discovery.wrap_anthropic(anthropic.Anthropic())
# or async:
client = discovery.wrap_anthropic(anthropic.AsyncAnthropic())

Google GenAI

model = discovery.wrap_google(genai.GenerativeModel("gemini-2.5-flash"))
response = model.generate_content("Hello")

Any Provider (generic trace)

result = discovery.trace(
    "My System", "OpenAI",
    lambda: openai.chat.completions.create(...),
    model="gpt-4",
    extract_usage=lambda r: {
        "input_tokens": r.usage.prompt_tokens,
        "output_tokens": r.usage.completion_tokens,
    },
)

Manual Reporting

discovery.capture(
    system_name="Resume Screener",
    provider="Anthropic",
    model="claude-sonnet-4-20250514",
    input_tokens=500,
    output_tokens=1500,
)

Configuration

v2.0.0 Migration from v1.0.0

v1.0.0 used monkey-patching (discovery.init()). v2.0.0 uses explicit wrapping:

# v1.0.0 (deprecated)
discovery.init()
client = anthropic.Anthropic()

# v2.0.0
client = discovery.wrap_anthropic(anthropic.Anthropic())

No global state is modified. Multiple discovery instances can coexist.