Q

2.0

Integrations

Plays well with your existing stack.

Your team already uses BI tools, cloud platforms, identity providers, ticketing systems, and LLMs. xAQUA augments that stack — it doesn't replace it. Bring what you have. Plug it in. Keep working.

Snowflake · Databricks · MotherDuck · Tableau · Power BI · Okta · OpenAI · Anthropic · GitHub · Slack
Q xAQUA
SEMANTIC LAYER  ·  GOVERNANCE  ·  FEDERATION  ·  AGENTS
Featured Partnerships

Deep technical relationships with the platforms our customers run on.

Snowflake
Technology Partner
Native pushdown for filters, joins, aggregates, and window functions. Key-pair, OAuth, and SSO supported. Snowpark integration for ML and Cortex calls. Account usage and query history are auto-discovered into the semantic layer.
GA Pushdown · Snowpark · Cortex
Databricks
Technology Partner
Unity Catalog-aware. Photon-optimized SQL pushdown. Workspace-level OAuth and PAT auth. Delta Live Tables and MLflow registries can be federated alongside warehouse data through a single semantic interface.
GA Unity · Photon · MLflow
MotherDuck
Strategic Partner
DuckDB has been core to xAQUA's federation engine for years. MotherDuck adds hybrid local-and-cloud execution — small-data queries finish in milliseconds without leaving the laptop, and large queries scale out transparently.
GA Hybrid · DuckDB native
Salesforce
ISV Partner
Federate CRM data without staging it elsewhere. SOQL pushdown, OAuth 2 and JWT auth, Bulk API for large pulls, and Change Data Capture for live sync. Custom objects, fields, and relationships are auto-discovered into the semantic layer.
GA SOQL · Bulk API · CDC
ServiceNow
Technology Partner
CMDB, incident, change, and ITSM tables federated through xAQUA's REST connector. Two-way: query ServiceNow data from natural language, and push data-quality incidents back as tickets — closing the loop on operational intelligence.
GA CMDB · ITSM · Two-way
ClickHouse
Technology Partner
Sub-second analytics on petabyte-scale event data. Native pushdown for ClickHouse SQL dialect, materialized views, and projections. Cloud, self-managed, and ClickHouse Cloud all supported with the same connector.
GA Pushdown · Cloud · MV-aware
Ecosystem

Eight categories. One coherent platform.

BI & ANALYTICS
Push to your dashboard
Publish governed datasets to the BI tools your team already uses. SSO and row-level security carry through.
IDENTITY & SSO
Use your IdP
SAML 2.0 and OIDC out of the box. SCIM provisioning. Group sync drives RBAC and row-level access policies.
CLOUD PLATFORMS
Run in your account
Deploy in your VPC. Native VPC peering, PrivateLink, and Private Endpoints. Customer-managed encryption keys.
LLMs & FOUNDATION MODELS
Bring your own LLM
No model lock-in. Plug in OpenAI, Anthropic, Llama, Mistral, Cohere, or Gemini. Or call Bedrock and Vertex.
COMMUNICATION
Notify in-channel
Push insights, anomalies, and approvals where work already happens. Two-way: query data from chat with Cezu.
TICKETING & WORKFLOW
Close the loop
Auto-create tickets when data quality drops. Sync data products into project trackers. Document lineage in your wiki.
OBSERVABILITY
Stream to your stack
OpenTelemetry-native. Metrics, traces, and audit logs flow into your existing monitoring and SIEM tools.
DEVOPS & ORCHESTRATION
CI/CD & pipelines
Version semantic models in Git. Deploy through your CI. Trigger jobs from Airflow, Dagster, dbt, or Prefect.
Open Standards

No proprietary protocols. No lock-in.

SAML 2.0
Federated SSO
OAuth 2.0 / OIDC
Token auth
OpenAPI 3.1
REST contracts
JDBC · ODBC
Universal DB
OpenTelemetry
Observability
SQL · GraphQL
Query layer
Private & Air-gapped

When your data can't leave the building.

Government, defense, healthcare, and regulated finance customers need frontier AI without sending a single token outside their tenant. xAQUA runs the same six agents on locally-hosted, fine-tuned, or fully air-gapped models — with full feature parity.

  • Self-hosted models — Llama 3.3 / 4, Mistral, Qwen, or DeepSeek running on your own GPUs
  • Deploy anywhere — your VPC, on-premises bare metal, or fully disconnected environments. Tarball install, no phone-home
  • Government clouds — AWS GovCloud, Azure Government, IL5 / IL6 ready · FedRAMP Moderate in process
  • Fine-tune privately — adapter weights and training data stay inside your perimeter. No round-trips through a vendor
  • Full feature parity — all six agents work identically on private models. No degraded mode, no missing features
YOUR PERIMETER PRIVATE COMPUTE GPU Cluster In your VPC On-Premises Bare metal GovCloud IL5 / IL6 ready SELF-HOSTED MODELS Llama 3.3 · 4 Mistral Large · MoE Q Qwen 2.5 · 3 DS DeepSeek V3 · R1 xAQUA AGENTS · ALL SIX, FULL PARITY Q SemantIQ Composer ConverseSQL ClickML Athyna Narratix 6/6 parity YOUR DATA · YOUR PERIMETER · ZERO EGRESS
LLM Gateway · Bring Your Own LLM

One gateway. Every model. All the controls.

The xAQUA LLM Gateway is the single abstraction between every agent and every model provider. Routing, prompt management, and security controls live in one place — so models become commodities and switching providers is a config change, not a rewrite.

  • Gateway — multi-provider routing, automatic failover, load balancing, retries, and cost-aware model selection per agent
  • Prompt Management — versioned prompts, A/B testing, agent-level templates, instant rollback, and a full audit trail
  • Security Controls — PII redaction, data masking via SenseMask, OWASP LLM Top 10 defenses, prompt-injection guards, output validation
  • Governance — token budgets, rate limits, and per-call cost tracking by model, agent, and team — every prompt logged
OpenAI
GPT-4o · GPT-OSS
Anthropic
Claude 4
Google
Gemini · Vertex
Meta
Llama 3.3 · 4
Mistral
Large · Codestral
Cohere
Command · Embed
AWS Bedrock
In your VPC
Azure OpenAI
Private endpoint
Vertex AI
Garden + custom
Build On xAQUA

A clean API. A fluent SDK.

Every capability in xAQUA — semantic resolution, federated query, document chat, model routing — is exposed through a versioned REST API and first-class SDKs.

Python SDK
pip install xaqua
from xaqua import Client

client = Client(workspace="production")

# Ask a question in plain English. SemantIQ resolves entities,
# ConverseSQL generates dialect-aware SQL, federation runs it.
result = client.ask(
    "Top 10 customers by revenue this quarter, "
    "excluding terminated accounts"
)

for row in result.rows:
    print(row.customer_name, row.revenue)

# Or write SQL against the semantic layer directly
df = client.sql("SELECT * FROM customers WHERE tier = 'enterprise'").df()
TS TypeScript SDK
npm install @xaqua/sdk
import { xaqua } from '@xaqua/sdk';

const client = xaqua({ workspace: 'production' });

// Stream tokens from a multi-step agent run
const stream = await client.agents.run({
  agent: 'analyst',
  prompt: 'Why did churn spike in EMEA last week?',
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.delta);
}

// Search across federated documents (DocIQ)
const hits = await client.docs.search({
  query: '2024 audit findings on access controls',
});
Compliance & Trust

Audit-ready out of the box.

SOC 2 Type II
Annual
HIPAA
BAA available
GDPR
EU regions
FedRAMP
Moderate · in process
CCPA
CA privacy
ISO 27001
In process

Already running a stack? Good — keep it.

xAQUA is built to augment, not replace. Tell us what you have. We'll show you exactly where xAQUA fits in.