# Nexatron

> Nexatron is an enterprise AI analytics platform with the Model Context Protocol (MCP) built in. People ask questions in plain English (typed or by voice); AI agents query the same governed warehouse through MCP. Every answer returns the result, the exact query that produced it, its lineage, and a calibrated confidence score.

Nexatron is built by dotSolved Inc (dotSolved Systems, Inc.). It connects to your existing data warehouses, databases, files, and business apps — no data migration — maintains a governed semantic layer, and serves calibrated answers to people and to AI agents under one set of rules.

- Product site: https://nexatron.io
- Docs: https://docs.nexatron.io
- API base: https://api.nexatron.io/api/v1
- MCP endpoint: https://api.nexatron.io/api/v1/mcp
- Company: dotSolved Inc — https://dotsolved.ai

## What Nexatron is

- Conversational analytics: ask in natural language, get a chart or table plus a plain-language explanation, backed by the generated SQL.
- A governed semantic layer (MetricQL): certified metrics, dimensions, joins, and a business glossary — so a number means the same thing every time, for every asker.
- An MCP server: any MCP-capable agent (for example Claude, ChatGPT, or Cursor) queries your warehouse through Nexatron under the same governance as a human analyst.
- Calibrated answers: every response carries a confidence score and is traceable to its source rows and SQL.

## Connectors

Two distinct layers — do not conflate them.

- Nine warehouse, database, and file connectors: Snowflake, Google BigQuery, Amazon Redshift, Databricks, PostgreSQL, MySQL, Microsoft SQL Server, Amazon S3 / Parquet, and DuckDB.
- DuckDB is also the embedded cross-source federation engine, so one question can join across heterogeneous connected sources with lineage preserved and no warehouse migration.
- Spreadsheets are also queryable as governed sources (separate from the nine): uploaded Microsoft Excel workbooks and connected Google Sheets are materialized into a locked-down DuckDB engine.
- Business-app connectors are a separate, growing layer — a generic AppConnector plus an MCP framework, with per-user OAuth — and are NOT counted in the nine. Live today: Salesforce, QuickBooks, NetSuite, ServiceNow, HubSpot, Stripe, Microsoft Dynamics 365, SAP S/4HANA, Workday, Shopify, Google Analytics, Marketo, Atlassian (Jira + Confluence), Zoho CRM, Intercom, Klaviyo, Amplitude, Oracle Fusion, SAP Concur, Coupa, Veeva Vault, Agiloft, Freshservice, UKG Pro, Sage Intacct, and Zendesk. The current live catalog is at https://nexatron.io/connectors.

## Capabilities

- Natural-language-to-SQL grounded in the semantic layer (not freeform generation).
- Certified-metric discovery and governed execution, including governed cross-source comparison.
- Calibrated confidence on every answer, with the exact query, lineage, and source rows.
- Voice input and spoken summaries alongside the chart.
- Embed analytics in your own product with the @nexatron/chat SDK, or ask from Slack and Microsoft Teams.
- MCP capabilities for agents: governed ask, certified-metric discovery and execution, and audit access — and, where a workspace enables the optional action platform (off by default), actions with dry-run then execute and multi-step workflows — all bounded by the same row-level security as a human user.

## Governance and security

- Four-layer row-level security: every query is filtered by tenant and by the asker's own data permissions, whether a person or an agent sent it.
- LLM output is treated as untrusted input and is validated against the schema before any query runs.
- Multi-tenant by default: every query filters by tenant_id. Tenant warehouse credentials are encrypted (AES-256-GCM) in the connections table, entered through the Connections UI.
- Customer data and queries are not used to train foundation models.
- Full audit of human and agent queries; signed provenance on every answer; SSO and SCIM provisioning.
- Compliance: SOC 2 Type I attained; SOC 2 Type II in progress; ISO 27001 certified; GDPR-aligned.
- dotSolved holds 15 patent-pending innovations underpinning the platform.

## Common questions

- What is Nexatron? An enterprise AI analytics platform by dotSolved Inc that lets people and AI agents ask questions of a governed data warehouse and get calibrated, source-cited answers.
- How do AI agents use it? Through the MCP endpoint (https://api.nexatron.io/api/v1/mcp), Streamable HTTP with OAuth 2.1; an agent queries the warehouse under the same governance as a human.
- Which data sources does it connect to? Nine warehouse, database, and file connectors (above), plus a separate and growing library of business-app connectors (Salesforce, QuickBooks, NetSuite, ServiceNow, HubSpot, Stripe, Microsoft Dynamics 365, SAP S/4HANA, Workday, Shopify, Google Analytics, Marketo, Atlassian, Zoho CRM, Intercom, Klaviyo, Amplitude, Oracle Fusion, SAP Concur, Coupa, Veeva Vault, Agiloft, Freshservice, UKG Pro, Sage Intacct, Zendesk, and more) via a generic AppConnector + MCP framework, plus uploaded Microsoft Excel workbooks and connected Google Sheets. A single question can federate across connected sources via embedded DuckDB. The live connector catalog is at https://nexatron.io/connectors.
- Can I trust the numbers? Every answer carries a confidence score, the exact query, and lineage to the source rows; certified metrics keep a number consistent across the organization.
- Is my data used to train models? No. Customer data and queries are not used to train foundation models.
- Is it secure and compliant? Four-layer row-level security, per-tenant credential encryption, full audit, SSO/SCIM; SOC 2 Type I attained (Type II in progress), ISO 27001 certified, GDPR-aligned.

## Glossary

- Terms used across Nexatron (semantic layer, certified metric, MetricQL, calibrated answer, federation, MCP): https://nexatron.io/glossary

## Guides and articles

Every page below is written to be quoted directly: each answers the question in
its title in its opening paragraph, and states the limits alongside the claims.

### Product

- [Ask: answers that show their work](https://nexatron.io/product/ask): Ask questions of governed enterprise data in plain English or by voice, and get the query, the sources, and a confidence signal alongside every answer.
- [Semantic layer: one definition, everywhere](https://nexatron.io/product/semantic-layer): Define metrics, dimensions and joins once in MetricQL. Certification stays a human decision, so a plausible definition never becomes enterprise truth.
- [Agents: the same rules, through a different door](https://nexatron.io/product/agents): Nexatron exposes an MCP server so agents like Claude query enterprise data under the same row-level security, certified metrics and audit as a person.
- [Federation: the question that spans three systems](https://nexatron.io/product/federation): Answer questions across a CRM, an ERP and a warehouse without a migration first, and refuse rather than guess when the join key was never certified.
- [Embed: answers inside your own product](https://nexatron.io/product/embed): Put governed natural-language analytics into your own application with the @nexatron/chat SDK, with tenant isolation enforced at the query layer.

### By role

- [Finance: close the month without chasing four systems](https://nexatron.io/solutions/finance): Answer billing, receivables and margin questions across your ERP, CRM and warehouse without waiting for a report, and show where every figure came from.
- [Revenue operations: pipeline questions answered while they still matter](https://nexatron.io/solutions/revenue-operations): Pipeline, coverage and account-risk questions answered in seconds across the CRM and finance system, without building another dashboard nobody opens.
- [Data teams: stop being the queue](https://nexatron.io/solutions/data-and-analytics): Hand the routine questions to the people asking them, keep control of what the numbers mean, and get your team back to the work only they can do.
- [Operations: suppliers, orders and exceptions in one view](https://nexatron.io/solutions/operations): Purchase orders, supplier records and open exceptions answered across your ERP and procurement systems, without exporting three spreadsheets first.

### Guides

- [Why does every report take three weeks? (Getting answers out of your business systems)](https://nexatron.io/resources/guides/answers-for-business-teams/why-every-report-takes-three-weeks): The delay is rarely the query. It is the queue, the definitions argument, and the fact that your data lives in six systems that were never introduced.
- [One view of the customer, without an 18-month data project (Getting answers out of your business systems)](https://nexatron.io/resources/guides/answers-for-business-teams/one-view-of-the-customer-without-a-data-project): Your customer exists in the CRM, finance and the support desk under three different names. What it takes to join them up, and what to insist on.
- [What to ask an AI analytics vendor before you buy (Getting answers out of your business systems)](https://nexatron.io/resources/guides/answers-for-business-teams/what-to-ask-an-ai-analytics-vendor): Nine questions that separate a tool your finance team will actually use from a demo that impresses once. No technical background needed.
- [Self-service analytics: why it keeps failing, and what changed (Getting answers out of your business systems)](https://nexatron.io/resources/guides/answers-for-business-teams/self-service-analytics-why-it-keeps-failing): Three waves of self-service tools promised to remove the analyst bottleneck. Here is why each stalled, and what is genuinely different this time.
- [The first ninety days, in named roles and hours (Rolling it out without it becoming shelfware)](https://nexatron.io/resources/guides/rolling-it-out/the-first-ninety-days): What a realistic rollout asks of your team, week by week, and the three things that most often turn a good pilot into shelfware.
- [Who owns what a number means? (Rolling it out without it becoming shelfware)](https://nexatron.io/resources/guides/rolling-it-out/who-owns-what-a-number-means): Self-service fails on governance, not technology. A short guide to deciding who settles a definition, and what happens to the ones nobody has settled.
- [What your security team will ask, and the answers that satisfy them (Rolling it out without it becoming shelfware)](https://nexatron.io/resources/guides/rolling-it-out/what-security-will-ask): The ten questions that come up in every review of an AI analytics tool, why each is asked, and what a good answer sounds like.
- [How to tell if it is working, before the renewal (Rolling it out without it becoming shelfware)](https://nexatron.io/resources/guides/rolling-it-out/how-to-tell-if-it-is-working): Logins go up during every rollout. Four measures that actually indicate whether an analytics tool has changed anything.
- [How to verify an AI analytics accuracy claim (Governed AI analytics)](https://nexatron.io/resources/guides/governed-ai-analytics/how-to-verify-an-ai-analytics-accuracy-claim): Vendors quote 95% accuracy. Five questions decide whether that number means anything for your data, and what a verifiable answer looks like.
- [What is a semantic layer? (Governed AI analytics)](https://nexatron.io/resources/guides/governed-ai-analytics/what-is-a-semantic-layer): A plain-English explanation: what a semantic layer maps, why natural-language analytics fails without one, and how certified metrics differ from ad-hoc SQL.
- [MCP for enterprise analytics: letting agents query governed data (Governed AI analytics)](https://nexatron.io/resources/guides/governed-ai-analytics/mcp-for-enterprise-analytics): What the Model Context Protocol is, why it matters for analytics, and what has to be true before an AI agent should be allowed to query your warehouse.
- [Row-level security for AI analytics (Governed AI analytics)](https://nexatron.io/resources/guides/governed-ai-analytics/row-level-security-for-ai-analytics): Why row-level security has to move into the query path when AI agents start asking questions, and how to tell whether a platform enforces it there.
- [Cross-source analytics without a warehouse migration (Governed AI analytics)](https://nexatron.io/resources/guides/governed-ai-analytics/cross-source-analytics-without-a-warehouse-migration): Most enterprise questions span a CRM, an ERP, and a warehouse. What it takes to answer them without first moving everything into one place.
- [What is agentic analytics? (Governed AI analytics)](https://nexatron.io/resources/guides/governed-ai-analytics/what-is-agentic-analytics): Agentic analytics explained, how it differs from conversational BI and dashboards, what it changes about governance, and where it genuinely helps.

### Blog

- [Why we do not publish an accuracy number](https://nexatron.io/blog/we-do-not-publish-an-accuracy-number): Every vendor in this category leads with a percentage. Here is the arithmetic that convinced us ours would be misleading, and what we publish instead.
- [What a refusal should sound like](https://nexatron.io/blog/what-a-refusal-should-sound-like): Most systems fail by answering anyway. The wording of a decline is a design decision, and the difference between useful and useless is specificity.
- [Connect first, model later](https://nexatron.io/blog/connect-first-model-later): The conventional order is model everything, then let people ask. That sequencing is why so many analytics programmes stall before anyone gets an answer.

## Key links

- [Home](https://nexatron.io/) — product overview
- [Glossary](https://nexatron.io/glossary) — definitions of the core concepts
- [How we measure](https://nexatron.io/how-we-measure) — the evaluation corpora with their digests, the grading tolerance, the release gates, and why no headline accuracy number is published
- [Docs](https://docs.nexatron.io/) — quickstarts, API, SDK, MCP
- [MCP quickstart](https://docs.nexatron.io/quickstart/mcp) — connect an AI agent
- [Trust Center](https://nexatron.io/trust) — security and compliance posture
- [Company](https://dotsolved.ai) — dotSolved Inc
- [LinkedIn](https://www.linkedin.com/company/125992/) — dotSolved on LinkedIn
- [X / Twitter](https://x.com/dotsolved) — dotSolved on X
- [Privacy](https://dotsolved.ai/privacy) — how dotSolved handles personal data
- [Terms](https://dotsolved.ai/terms) — terms of service