# Omnia vs Relevance AI

> Omnia builds, tests and gates agents inside your Kubernetes cluster, redacts personal data before it’s stored, rolls versions out in steps and enforces policy on every tool call. Relevance AI does its testing and gating on its own multi-tenant cloud.

Category: Agent platforms. Checked against Relevance AI’s published material on 2026-10-22. Canonical page: https://www.altairalabs.ai/vs/relevance-ai/

## Relevance and Omnia run the same build, test and release loop, and Omnia runs it in your own cluster.

Relevance builds agents, tests them on simulated users and can block a failing publish, on its own multi-tenant cloud. On Omnia the conversations stay in your cluster, with personal data removed before they’re stored.

## When it matters: A customer reads their card number out to the agent.

**On Omnia:** The number is redacted before the session is recorded, redacted again before anything reaches long-term memory, and replaced in tool responses before the model sees it. What’s left is stored in your cluster under your retention rules. (Omnia docs: https://omnia.altairalabs.ai/how-to/privacy/configure-privacy-policies/)

**On Relevance AI:** Relevance runs a multi-tenant architecture with customer data logically separated, in a US, EU or Australian region. Single-tenant is described as in the works. (source: https://relevanceai.com/docs/admin/security)

## The data stays inside your boundary

Omnia installs into your Kubernetes cluster. Sessions move from Redis to Postgres to object storage you own, and the archive is Parquet you query from Snowflake, BigQuery or DuckDB. Traces go live to your OpenTelemetry collector. (Omnia docs: https://omnia.altairalabs.ai/explanation/agents/sessions/)

Relevance AI: Data is stored on Relevance’s platform; telemetry can be exported in OpenTelemetry format to your S3 bucket. (source: https://relevanceai.com/docs/enterprise/streaming-events.md)

## A bad version comes back out on its own

A version that passes the gate starts on part of the traffic and gets more while the metrics hold; if they drop, the previous version takes the traffic back. Someone mid-conversation stays on the version they started with. *(enterprise)* (Omnia docs: https://promptarena.altairalabs.ai/arena/how-to/interfaces/run-in-ci/)

Relevance AI: Restoring a version creates a new draft, which you then publish. Stepped rollout isn’t documented. (source: https://relevanceai.com/docs/build/agents/build-your-agent/version-history.md)

## Policy that knows who’s asking

The caller’s identity travels through the agent to every tool. A CEL rule compares each call’s arguments with it, and a refund to an account the caller doesn’t own is refused. *(enterprise)* (Omnia docs: https://omnia.altairalabs.ai/explanation/platform/delegation-tokens/)

Relevance AI: Human approval per tool. Argument-level policy and end-user identity in tool calls aren’t documented. (source: https://relevanceai.com/docs/build/workforces/workforce-features/approvals-and-escalations.md)

## At a glance

| Topic | Omnia | Omnia docs | Relevance AI | Relevance AI docs | Source |
|---|---|---|---|---|---|
| Build from a conversation | The Arena builder writes the agent and its tests from a brief, and runs them on real providers. *(enterprise)* | https://omnia.altairalabs.ai/how-to/evaluation/generate-a-project-with-the-builder/ | Invent builds and manages agents from a conversation. | documented | https://relevanceai.com/docs/build/invent/invent.md |
| Runs in your infrastructure | Your cluster, or AgentCore, Vertex or Foundry in your account. | https://promptarena.altairalabs.ai/arena/explanation/deploy/overview/ | Multi-tenant SaaS. | not documented | https://relevanceai.com/docs/admin/security |
| Simulated users | Multi-turn, against real tools, adversarial included. | https://promptarena.altairalabs.ai/arena/explanation/scenario-design/ | Multi-turn, with live or simulated tools. | documented | https://relevanceai.com/docs/build/agents/evals/test-sets.md |
| Eval-gated release | The suite decides whether a version ships. | https://promptarena.altairalabs.ai/arena/how-to/interfaces/run-in-ci/ | Optional publish block, Enterprise first. | documented | https://relevanceai.com/docs/build/agents/evals/introduction.md |
| Voice testing | Barge-in, overlapping speech and dead air. | https://promptarena.altairalabs.ai/arena/how-to/voice/setup-voice-testing/ | Phone agents ship; testing them isn’t documented. | not documented | https://relevanceai.com/docs/build/agents/build-your-agent/agent-modes/phone-calls.md |
| Stepped rollout, auto rollback | Traffic shifts in steps and rolls back on its own. | https://omnia.altairalabs.ai/tutorials/progressive-rollouts/ | Not documented. | not documented | https://relevanceai.com/docs/build/agents/build-your-agent/version-history.md |
| Live telemetry to your collector | OpenTelemetry and Prometheus. | https://omnia.altairalabs.ai/reference/platform/telemetry/ | OpenTelemetry events to S3. | partly documented | https://relevanceai.com/docs/enterprise/streaming-events.md |
| Argument-level tool policy | A CEL rule on every call, before the tool runs. *(enterprise)* | https://omnia.altairalabs.ai/how-to/security/configure-tool-policies/ | Not documented. | not documented | https://relevanceai.com/docs/build/workforces/workforce-features/approvals-and-escalations.md |
| Any model, including your own | Seven native protocols and any OpenAI-compatible URL. | https://omnia.altairalabs.ai/reference/core/provider/ | OpenAI, Gemini, Azure, OpenRouter; Anthropic via your cloud account. | partly documented | https://relevanceai.com/docs/integrations/llm-integrations/anthropic.md |

## Questions

**Can I self-host Relevance AI?**

Its security documentation describes a multi-tenant platform with regional storage, and single-tenant as in the works. Omnia runs in your own cluster from day one.

**Do both test agents with simulated users?**

Yes. Omnia adds adversarial scenarios and duplex voice testing, and runs them inside your boundary against your real tools.

**What does Omnia cost?**

It’s free while you build and the core is free in production. The enterprise licence is one price per environment per year from £30,000, and nothing is metered.