The customer-resident agent
The agent runs inside your network and is the only DataNivra component that reads your data. Each release on Downloads provides:
| What | Use it for | Manual |
|---|---|---|
Signed multi-architecture image ghcr.io/manuelbomi/datanivra-agent@sha256:… (linux/amd64, linux/arm64) | every install method; deploy it by digest | Agent installation |
| Helm chart and values file | Kubernetes, with a restricted security context and NetworkPolicy | Agent installation |
| Docker, Docker Compose, kustomize, Azure VNet and AWS VPC packages | one-command and infrastructure-as-code installs | Agent installation |
SHA256SUMS + cosign bundle, SBOM, release notes | verifying what you deploy | Agent installation → Verify |
| Documentation package | offline copies of the install runbooks | included in each release |
Optional: Isolated execution runs every row-level stage in a fresh container, gVisor sandbox or Firecracker microVM.
Command line, SDKs and MCP
All client packages are released together and follow Semantic Versioning; the contract packages carry the API version (v1 is 1.x).
| Package | Install | Manual |
|---|---|---|
CLI datanivra-cli (Python 3.12 or 3.13) | pipx install datanivra-cli | CLI |
Python SDK datanivra-sdk | pip install datanivra-sdk | SDKs and OpenAPI, SDK helpers |
TypeScript SDK @datanivra/sdk | npm install @datanivra/sdk | SDKs and OpenAPI, SDK helpers |
Contracts datanivra-contracts / @datanivra/contracts | installed with the SDKs | API overview |
MCP adapter datanivra-mcp (stdio) | pipx install datanivra-mcp | Test data for AI agents (MCP) |
| Hosted MCP endpoint | no install: https://api.datanivra.com/mcp with a token | Test data for AI agents (MCP) |
| OpenAPI 3.1 document | GET /openapi.json on the API, or the API reference | SDKs and OpenAPI |
Checksums of each client release are published under Downloads.
Industry packs, examples and CI recipes
| What | Where | Manual |
|---|---|---|
| Industry pack bundles (schemas, scenarios, policy templates, sample synthetic data) | each pack's page under Industry Packs | the pack page itself |
| Public examples: CI pipelines, test-data API calls, MCP setups, integrations (MIT) | github.com/DataNivra/datanivra-examples and Examples | each example's README |
| CI recipes (GitHub Actions, GitLab CI, Azure DevOps, Jenkins, CircleCI) | CI recipes | same page |
| Test-data profiles | Test-data profiles | same page |
Which manual for which task
| Task | Start with |
|---|---|
| First steps, end to end | Getting started |
| Install, verify, upgrade an agent | Agent installation |
| Require stronger isolation of row-level work | Isolated execution |
| Automate requests from code or CI | Universal Test Data API, CLI, CI recipes |
| Connect an AI agent | Test data for AI agents (MCP) |
| Operate and troubleshoot | Operator runbooks, Troubleshooting, support center |
| Security review | Agent protocol, Security, Known limitations |