8+
CI/CD platforms
12+
LLM providers
SaaS / VPC / On-prem
Deployment options
REST API + Python SDK
Developer access
Webhooks
Async callback support
Drop into your existing pipeline in minutes
Don't see yours? We support any CI/CD tool via REST API and webhook callbacks.
- GitHub Actions: YAML action in GitHub Marketplace; add qapitol/qave-action@v2
- Jenkins: Jenkins plugin + declarative pipeline step
- Azure DevOps: Extension in Azure Marketplace for Pipelines YAML
- GitLab CI: Native .gitlab-ci.yml integration
- CircleCI: Qapitol orb on CircleCI Orb Registry
- Bitbucket Pipelines: Pipe for PR, push, scheduled triggers
- Harness: Step library integration for CD pipelines
- TeamCity: Plugin with threshold-based conditions
Test the models you're actually using
LLM compatibility matrix covering 12+ providers with evaluation, red teaming, and compliance support per model family.
- OpenAI: GPT-4o, GPT-4 Turbo, GPT-3.5, o1, o3
- Anthropic: Claude 3.5 Sonnet, Claude 3 Opus, Haiku
- Google: Gemini 1.5 Pro, Flash, Ultra
- Meta: Llama 3.1 70B, 8B, Llama 2
- Mistral: Mistral Large, 7B, Mixtral 8x7B
- Cohere: Command R+, Command R
- Azure OpenAI: GPT-4o, GPT-4 (Azure-hosted)
- AWS Bedrock: Claude, Llama, Titan, Mistral
- Custom/Fine-tuned: ONNX, HuggingFace, API-served
- Local/Air-gapped: Ollama, vLLM, LM Studio
Your data perimeter. Your deployment choice.
Three deployment modes: SaaS (Qapitol Cloud), VPC Deployment, and On-Premise / Air-Gap.
- SaaS · Qapitol Cloud (Most Common): API key onboarding, live in under 15 minutes; SOC 2 compliant infrastructure; 99.9% SLA with automatic failover; shared multi-tenant environment; ideal for startups and mid-market teams
- VPC Deployment: deploy inside AWS, Azure, or GCP; data never crosses perimeter; bring your own KMS encryption keys; dedicated tenant, no data commingling; ideal for BFSI, healthcare, enterprise
- On-Premise / Air-Gap: fully isolated, no internet connectivity; deploy to data centre or sovereign cloud; includes hardware sizing and architecture review; 4-week onboarding engagement; ideal for Government, Defence, regulated industries
One function call to add AI quality gates
Python SDK demonstration using QAVEClient with evaluation checks: hallucination, toxicity, pii_exposure, policy_compliance.
- Python SDK: pip install qapitol-qave · Python 3.9+
- JavaScript/TypeScript SDK: npm install @qapitol/qave
- REST API: OpenAPI 3.1 spec, Bearer token auth
- Webhooks: Async eval callbacks, HMAC-SHA256 signed, retry on failure
AI quality metrics in the dashboards you already use
Observability and monitoring integrations.
- Datadog: Custom metrics, pre-built dashboards, threshold alerting
- Grafana: Native dashboard templates, Prometheus data source
- Prometheus: OpenMetrics format endpoint
- Splunk: Log forwarding via Splunk HEC
- Microsoft Sentinel: SIEM log integration
- PagerDuty: Alert routing for threshold breaches
Buy through your cloud provider's marketplace
Cloud marketplace availability.
- AWS Marketplace: Live, available now
- Azure Marketplace: Listing in review
- GCP Marketplace: Coming Q3 2025