Skip to main content

Proofbeak

Check whether a service meets your requirements. Decide with evidence.

English · Türkçe · PyPI

Proofbeak is a self-hosted acceptance-testing and evidence service for developers and AI agents. Define the behavior your application needs, choose candidate services, and run the same scenario against each one. Proofbeak reports what it observed so you can make an informed service-selection or integration decision.

You write your own scenarios in JSON by combining available connector operations, test data and acceptance conditions. The included examples are starting points you can adapt.

How it works

  1. Define the preparation, actions, observations and expected results.
  2. Choose compatible services from the configured candidates.
  3. Validate the definition before starting provider work.
  4. Run the trial and check its outcome through independent reads.
  5. Inspect the comparison, evidence and separate cleanup status.
Result Meaning
pass The observations satisfy the requested conditions.
fail An observed value violates a requested condition.
inconclusive Missing observations or an operational issue prevent a conclusion.

JSON and HTML reports include expected and observed values, before/after records and evidence provenance. The developer or external agent makes the final choice.

Quick start

Use Python 3.11+ on Linux or macOS. Install from PyPI in a virtual environment; the local example needs no provider account or model service:

python3 -m venv .venv
. .venv/bin/activate
python -m pip install proofbeak==0.7.0
proofbeak example --output scenario.json
proofbeak validate scenario.json \
  --candidate scenario:filesystem
proofbeak run scenario.json \
  --candidate scenario:filesystem --key first-trial \
  --output reports/first-trial.json --html reports/first-trial.html

Open reports/first-trial.html to inspect the result. The example uses disposable files. Copy the JSON and edit it for your condition; use a new key for a changed definition. Repeating the same request and key retrieves the original run.

Follow the scenario guide to author a definition and apply it to multiple configured services.

Interfaces and scope

Use Proofbeak through CLI, REST or MCP. Provider connections and credentials are configured by the operator; the external agent runtime stays under your control.

Compose scenarios over configured REST APIs, MCP tools, SQLite/PostgreSQL transactions, Redis Streams and HTTP search/model services. Connectors publish input/output schemas; scenarios can include bounded polling, output-schema and ranking checks. Follow the capability guide for setup and operating limits. New scenarios use JSON; new protocols require a trusted connector.

Run directly in Python, in a container or on standard Kubernetes/Helm. The current deployment scope uses one coordinator with SQLite state. See the deployment guide for configuration and operating limits.

Documentation

Contributing · Architecture · Roadmap · Apache-2.0

Metadata

Release files for proofbeak 0.7.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for proofbeak 0.7.0
File Size Uploaded Provenance Actions
proofbeak-0.7.0.tar.gz 265.0 kB With provenance Details for proofbeak-0.7.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for proofbeak 0.7.0
File Interpreter ABI Platform Actions
proofbeak-0.7.0-py3-none-any.whl Python 3 none any Details for proofbeak-0.7.0-py3-none-any.whl

Total release size: 368.4 kB

Release files / proofbeak-0.7.0.tar.gz

Download URL proofbeak-0.7.0.tar.gz
Size 265.0 kB
Tags Source
SHA-256 checksum
How to use checksums
9cf41a57293e7e03c3a531fb8ac52a9bc8bc4696351ca6e8d8e7d7ae63298479
BLAKE2b-256 checksum
How to use checksums
d8f74a96c38e61ac62d3ace1dca3433c84e94c2c83b860e61009907d32268a7e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 13, 2026.

Transparency log

Release files / proofbeak-0.7.0-py3-none-any.whl

Download URL proofbeak-0.7.0-py3-none-any.whl
Size 103.4 kB
Tags Python 3
SHA-256 checksum
How to use checksums
65f8b5f2123a1f60f56cf45380348f1bda3ee0541d88376bc55542b6a2c78583
BLAKE2b-256 checksum
How to use checksums
e16fb260fc77f80862c6fd1a3ff2b31c053c768afd93836c79beda25929edaf8
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 13, 2026.

Transparency log

Security for proofbeak 0.7.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for proofbeak-0.7.0.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Nov 13, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page