Skip to main content

Python SDK for Redraven (fireraven.ai) — generate, run-against-client-LLM, and evaluate red-team tests.

Project description

redraven (Python SDK)

The official Python SDK for Redraven — the LLM red-teaming and policy-evaluation platform from fireraven.ai.

Redraven generates adversarial prompts, evaluates your LLM's responses against your safety policies and certifications, and surfaces failures so you can harden your AI products.

This SDK lets you drive the full Redraven flow from your own Python code:

  1. Generate (or pick) a test dataset on Redraven.
  2. Run each test case locally through your own LLM — your model API key never leaves your process.
  3. Submit responses back to Redraven for scoring; receive a summary you can act on.

The SDK talks only to public endpoints on the Redraven backend, authenticated with your organization API key (issued from the Redraven app).

Platforms

Installation

Install from PyPI:

uv add redraven
# or
pip install redraven

uv is preferred for reproducible installs.

Quick start

import asyncio
import redraven

async def my_llm(prompt: str) -> str:
    # Call your own LLM here. Your key stays in your process.
    return f"echo: {prompt}"

async def main():
    async with redraven.Client() as client:  # reads REDRAVEN_API_KEY + REDRAVEN_BASE_URL
        result = await client.run_test(
            test_id="<your-existing-test-id>",
            llm=my_llm,
            concurrency=4,
        )
        print(result.model_dump())

asyncio.run(main())

Core methods

  1. generate_test(generate_kwargs, wait_for_dataset=False, ...) -> str (returns test_id)
  2. wait_for_dataset_ready(test_id, ...) -> None
  3. run_test(test_id, llm, ...) -> EvalSummary
  4. generate_and_run_test(generate_kwargs, llm, ...) -> EvalSummary

Documentation

For private contributor documentation, use the repository internal docs.

License

Proprietary — © fireraven.ai. Contact us if you'd like to integrate Redraven into your product.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

redraven-0.1.2.tar.gz (39.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

redraven-0.1.2-py3-none-any.whl (10.5 kB view details)

Uploaded Python 3

File details

Details for the file redraven-0.1.2.tar.gz.

File metadata

  • Download URL: redraven-0.1.2.tar.gz
  • Upload date:
  • Size: 39.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: uv/0.9.30 {"installer":{"name":"uv","version":"0.9.30","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Debian GNU/Linux","version":"12","id":"bookworm","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for redraven-0.1.2.tar.gz
Algorithm Hash digest
SHA256 44afe69a0e66c4669efd7c64d7d460aaac345fecc601d51b4cd74887dda2d9ad
MD5 a493e33ee2c76c1bbcf29d041ba891b3
BLAKE2b-256 0210fecb0ba86ba05531490ee59a17cbfd311663b3c8fcb6aa021f21647e020f

See more details on using hashes here.

File details

Details for the file redraven-0.1.2-py3-none-any.whl.

File metadata

  • Download URL: redraven-0.1.2-py3-none-any.whl
  • Upload date:
  • Size: 10.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: uv/0.9.30 {"installer":{"name":"uv","version":"0.9.30","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Debian GNU/Linux","version":"12","id":"bookworm","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for redraven-0.1.2-py3-none-any.whl
Algorithm Hash digest
SHA256 136792f011b0422df257354f04d240967db457fbc9359bb0c0c061e352626a08
MD5 2aad1c1aeaccde4cdf07e409ad4129c3
BLAKE2b-256 df97968bf0b167dde1c29814fa5705608baa634cd603a8e94a9f705742db8177

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page