Commercient

Commercient Data Lake / Data Hub

Your business data, ready for apps and AI — on the database you already trust

A backend-as-a-service for Microsoft SQL Server. Bring your ERP, CRM, e-commerce, and database data into one governed, always-current SQL Server database per customer — and get instant REST and GraphQL APIs, AI-agent access over MCP, and a developer CLI on top of it. No pipelines to build, no API layer to write.

Start free from the terminal Talk to us

What is Commercient Data Lake?

One governed home for all your company data — built to be used, not just stored.

Most companies run their business across an ERP, a CRM, an e-commerce platform, and a handful of databases — and every new app, report, or AI project starts with the same grind: getting to the data. Commercient Data Lake removes that step. It gives every customer their own managed Microsoft SQL Server database, kept current from the systems they already run, and puts three ready-made surfaces on top of it.

An API surface — expose exactly the tables, views, and procedures you choose, and REST and GraphQL endpoints exist immediately, secured by API keys scoped down to entity and field level and enforced in the database itself. An AI surface — Claude and other MCP-capable agents connect with a revocable key and can query your data or, on the governed admin plane, operate the platform; bundled agent skills teach them how. A developer surface — the dlake CLI drives everything from sign-up to schema, queries, exports, and administration, ready for scripts and CI.

Because it is real SQL Server underneath, the platform's guarantees are database guarantees: row-level security, auditing, time travel, and role-based permissions are native capabilities — not features bolted onto a proprietary store. Your team's SQL skills and tooling carry straight over.

Who it's for

Three ways teams use it from day one.

Teams building on company data

Ship an internal tool, customer portal, or dashboard against live ERP and CRM data — without asking IT for a database, an API layer, or a sync pipeline first.

Companies enabling AI agents

Give AI agents governed access to business data: scoped keys, database-enforced permissions, full audit. Let agents answer from all your systems instead of one.

Commercient SYNC customers

Already syncing ERP to CRM? The Data Lake adds an API and AI layer over the same connected data — more than 150 systems reachable through SYNC.

What you get

The pieces of a production backend, present on day one.

Instant APIs

REST and GraphQL endpoints over exactly the entities you expose — with keys scoped to entity and field level, enforced in the database.

AI-agent ready

MCP connectors for any capable agent: a data plane for querying, a governed admin plane for operating the platform, and bundled skills that teach the agent both.

Your systems, connected

SYNC reaches 150+ ERPs, CRMs, and e-commerce systems; built-in connectors cover HubSpot, Stripe, Salesforce, ServiceTitan, SQL Server, and ODBC; files ingest as CSV, Parquet, or XML.

A real developer CLI

dlake — sign-up, schema, queries, exports, keys, and the whole admin surface, scriptable and CI-friendly.

Three ways to use it

Pick your surface — they all see the same governed data.

AI agents — Claude Code & MCP

Connect Claude Code with one command, or add a custom connector in Claude — then query your data and manage the whole platform in plain language, under scoped, revocable keys. How to connect →

The dlake CLI

Sign up, build schema, query, export, and administer everything from the terminal — scriptable, CI-friendly, one self-contained binary. Commands & quickstart →

REST & GraphQL APIs

Expose the entities you choose and endpoints exist immediately — plus events and signed webhooks for downstream systems. How access works →

How it works

From nothing to a queryable, API-served backend in three steps.

  1. Sign up from the terminal. dlake register start creates your tenant; verify by email and your Data Lake is provisioned and seeded automatically — the first status check after seeding hands you a working API key.
  2. Connect your systems. Pick your CRM and ERP in the setup wizard — or start empty and ingest files.
  3. Build. Point your app at the REST/GraphQL API, your AI agent at the MCP connector, or your scripts at the CLI.

Governed by the database, not by promises

Security is enforced where the data lives.

Per-customer isolationEvery tenant gets its own SQL Server database — separation is structural, not logical.
Keys, scoped and revocableAPI keys can be limited to entities and fields, and revoking one takes effect immediately.
Row-level securityFine-grained visibility rules run inside SQL Server on every query, whoever asks.
Everything auditedAdministrative actions and data changes leave an audit trail you can query like any other data.

Getting help

The full product guide travels with your tenant — reachable from every surface.

In the console

The web console's Help section carries the complete product guide for your tenant.

From the CLI

dlake guide help, dlake guide api, and dlake guide cli print the guides; dlake skills show <name> opens the task playbooks. Every command answers --help.

For AI agents

Agents fetch the same content live over MCP: get_help_docs, get_api_guide, and the skills via list_skills / get_skill — always current, no update needed.

Questions

Is there a backend-as-a-service for SQL Server?

Yes — Commercient Data Lake provides a managed SQL Server database per customer with instant REST and GraphQL APIs, row-level security, and AI-agent access over MCP.

Can AI agents access my ERP and CRM data safely?

Yes — agents connect over MCP with scoped, revocable API keys; permissions are enforced in the database, and every admin action is audited.

Do I need to migrate off my current systems?

No — SYNC keeps the Data Lake current from the systems you already run, bidirectionally where supported.

Is there a free way to try it?

Yes — self-service sign-up from the CLI provisions a working tenant with a temporary owner key: npm install -g datalake, then dlake register start.