schemathesis

Catch API bugs before your users do.

Schemathesis tests OpenAPI and GraphQL APIs by generating inputs from your schema, adapting to server responses, and chaining operations into realistic workflows.

Finding bugs that manual testing missed

Verdict, timeline and every failure in one offline page: --report html

# Test a demo API - finds real bugs in 30 seconds
uvx schemathesis run https://example.schemathesis.io/openapi.json

# Test your own API
uvx schemathesis run https://your-api.com/openapi.json

Schemathesis sends real requests that create, change and delete data. Point it at a dev or staging environment, not production.

• 💥 500 errors that crash your API on edge case inputs
• 📋 Schema violations where your API returns different data than documented
• 🚪 Validation bypasses where invalid data gets accepted
• 🔗 Integration failures when responses don't match client expectations
• 🔄 Stateful bugs where operations work individually but fail in realistic workflows
• 🕵️ Broken object-level authorization where one user can read another user's objects

• ⚙️ Config file - auth, phases, and per-operation overrides in schemathesis.toml. No Python.
• 🔐 Authentication - static headers, Basic, per-security-scheme credentials, or custom refresh logic.
• 🔗 Stateful testing - operation links inferred from your OpenAPI schema, or the GraphQL type graph, no manual wiring.
• 🧠 Adaptive testing - learns constraints, ids, and auth from responses, reusing them mid-run.
• ✅ Custom checks - assert your own business rules next to the built-in ones.
• 📖 Fuzz dictionaries - mix real ids, wordlists, or LLM-generated payloads into generated data.
• 🐌 Rate limiting - cap the request rate, or use auto to follow Retry-After on 429.
• 📊 Reports - an offline HTML page with every failure and its curl command, plus JUnit, VCR, HAR, NDJSON, JSON, and Allure.
• 🎯 Schema coverage - keyword-level coverage report showing which constraints your tests exercised.
• 🔁 Replay & baseline - re-run past failures and fail CI only on new ones.

Command Line:

uv tool install schemathesis
schemathesis run https://your-api.com/openapi.json

Config file (schemathesis.toml, no Python needed):

headers = { Authorization = "Bearer ${API_TOKEN}" }
generation.max-examples = 500
rate-limit = "auto"

Python Tests:

uv add --dev schemathesis

import schemathesis

schema = schemathesis.openapi.from_url("https://your-api.com/openapi.json")


@schema.parametrize()
def test_api(case):
    # Tests with random data, edge cases, and invalid inputs
    case.call_and_validate()


# Stateful testing: Tests workflows like: create user -> get user -> delete user
APIWorkflow = schema.as_state_machine()
# Creates a test class for pytest/unittest
TestAPI = APIWorkflow.TestCase

CI/CD:

- uses: schemathesis/action@v3
  with:
    schema: "https://your-api.com/openapi.json"

Used by teams at Spotify, WordPress, JetBrains, and Red Hat.

"Schemathesis is the best tool for fuzz testing of REST APIs on the market. We at Red Hat use it for examining our applications in functional and integration testing levels." - Dmitry Misharov, RedHat

🏆 Trophy case: crashes and schema violations Schemathesis found in vLLM, Qdrant, OpenObserve, Ory Kratos and Formbricks, each with a link to the fix.

🔬 Live Benchmarks: Schemathesis runs for an hour against real-world APIs such as Apache Airflow, Open WebUI and Prefect. For each one:

• Code coverage and API schema coverage: operations, parameters, schema keywords (schema coverage measured with TraceCov)
• Issues found, by category
• How coverage grows over the run, phase by phase

📚 Documentation with guides, examples, and API reference. Start here:

• Quick Start - 5 minutes
• CLI Tutorial - 20 minutes
• Pytest Tutorial - 15 minutes

⚠️ Upgrading from older versions? Check our Migration Guide for key changes.

• 💬 Discord community
• 🐛 GitHub issues

We welcome contributions! See our contributing guidelines and join discussions in issues or Discord.

Schemathesis is built on top of Hypothesis, a powerful property-based testing library for Python.

This project is licensed under the terms of the MIT license.