Installation

Install QWED SDKs for Python, TypeScript, Go, or Rust. Includes self-hosted full stack setup instructions for on-premise deployments.

Get QWED up and running in minutes.

​

Python SDK

pip install qwed

After installing, run the onboarding wizard to set up verification engines, configure your LLM provider, and generate a local API key:

qwed init

The wizard runs four checks. First, it confirms that core engines (SymPy, Z3, AST, SQLGlot) are installed. Next, it walks you through selecting a provider — NVIDIA NIM, OpenAI, Anthropic, Google Gemini, or any OpenAI-compatible endpoint. It then validates your credentials. Finally, it bootstraps a ready-to-use API key. For CI pipelines, use --non-interactive. See the CLI reference for all options.

After init, verify your setup and run the built-in test suite:

qwed doctor  # system health check — engines, provider, server, database
qwed test    # 12 deterministic tests — all must pass before production

qwed doctor reports the status of verification engines, LLM provider connectivity, and the database. qwed test runs deterministic tests across Math, Logic, SQL, and Code engines to confirm everything works correctly. See the CLI reference for details on both commands.

​

Requirements

• Python 3.10+
• Optional: Redis (for caching)

​

Symbolic engine (optional)

From v8.0.0, the Symbolic engine runs CrossHair only inside a Docker sandbox. Install the symbolic extra, which includes CrossHair and the docker client, then build the sandbox image from a qwed-verification checkout:

pip install "qwed[symbolic]"
docker build -t qwed-crosshair-sandbox:0.0.101 docker/crosshair-sandbox

The image comes from docker/crosshair-sandbox/Dockerfile and is never pulled at runtime. To use a different tag, set QWED_CROSSHAIR_IMAGE. Without Docker or the image, symbolic verification returns BLOCKED. See Docker sandbox requirement.

​

TypeScript SDK

npm install @qwed-ai/sdk
# or
yarn add @qwed-ai/sdk
# or
pnpm add @qwed-ai/sdk

​

Go SDK

go get github.com/QWED-AI/qwed-verification/sdk-go/v9@v9.1.0

​

Rust SDK

# Cargo.toml
[dependencies]
qwed = "9.1.0"

---

​

Full stack (self-hosted)

To run the complete QWED stack locally (API server, Postgres, Redis, Jaeger, Prometheus, and Grafana), use Docker Compose:

# Clone the repository
git clone https://github.com/QWED-AI/qwed-verification.git
cd qwed-verification

# Copy the settings template, then fill in every required value
cp deploy/.env.example deploy/.env

# Build and start the stack from the repository root
docker compose -f deploy/docker-compose.yml up -d

Compose refuses to start while a required value in deploy/.env is empty. See Self-hosting for the full list.

To run only the API server from source instead:

pip install -e ".[server]"
python -m uvicorn qwed_new.api.main:app --reload

​

Docker image

The qwedai/qwed-verification image on Docker Hub runs the API server on port 8000. Pass the required secrets from your shell when you start it:

# Generate each secret once and keep it. Changing the lookup or audit key later
# invalidates existing API keys or audit signatures.
for name in API_KEY_SECRET QWED_JWT_SECRET_KEY QWED_API_KEY_LOOKUP_SECRET QWED_AUDIT_SECRET_KEY; do
  export "$name=$(python -c 'import secrets; print(secrets.token_urlsafe(48))')"
done

docker run -p 8000:8000 \
  -v qwed-data:/var/lib/qwed \
  -e QWED_CORS_ORIGINS="http://localhost:3000" \
  -e API_KEY_SECRET \
  -e QWED_JWT_SECRET_KEY \
  -e QWED_API_KEY_LOOKUP_SECRET \
  -e QWED_AUDIT_SECRET_KEY \
  qwedai/qwed-verification:9.1.0

The server names no model. For requests that use a provider (natural-language, logic, statistics, and fact), also pass the provider’s settings, for example -e ACTIVE_PROVIDER -e OPENAI_API_KEY -e OPENAI_MODEL. The container doesn’t read the qwed init settings on your host.

The image runs the API server from version 8.0.0. Earlier tags ran the QWED GitHub Action. To scan a repository, use the GitHub Action instead of the image.

​

Infrastructure URLs

Service	URL
API	http://localhost:8000
Grafana	http://localhost:3000
Prometheus	http://localhost:9090
Jaeger	http://localhost:16686

Enterprise support coming soon: Managed hosting and dedicated support. Contact [e-mail]

---

​

Next steps

• Quick start guide
• Core concepts