Reliable answers
Every answer comes from a documented table, with its source.
Myriade documents your data, learns your definitions and enforces access rules for you. Your data team reviews & validates.
https://app.myriade.ai/mcp On Team and Enterprise plans, an Owner adds it in the organisation settings.
Full guideBusiness, Enterprise and Edu plans, on the web. On Business, an admin or owner turns the feature on.
Full guideOnly admins add custom connectors. On Free, Pro and Student plans, the account owner is the admin.
Full guideGitHub Copilot in VS Code: see Coding agents.
Full guidegemini mcp add --transport http myriade https://app.myriade.ai/mcp Gemini Enterprise: an admin adds the MCP address in the Google Cloud console.
Full guidenpx add-mcp https://app.myriade.ai/mcp You had 18,240 active customers in October, up 4.2% on September.
Active means 2 orders in the last 6 months.
We started Myriade 4 years ago, trying to build the most advanced AI analytics platform in the world.
Our dream: make data accessible, easy to use and impactful.
The problem: AI on data works in demos, not in real enterprise contexts.
To achieve this dream, we had to build the foundations that make AI reliable on real, messy, enterprise data.
So we built the data governance platform for the AI era.
We love it. Our users love it. We think you will too.

Let’s face it: most data catalogs are empty, and nobody opens them. We built ours to be written by AI agents, for AI agents. It documents every table and column, and asks your data team only when it can’t tell.
The result: a catalog that is complete, correct and kept up to date, with no effort from your team.
We documented 100% of our tables in under a week with Myriade.
Building a semantic layer is complex, boring and never finished. We made it simple.
Myriade learns from your usage, your concepts, your data, and suggests definitions. You review & validate.
The result: one set of definitions & KPIs that every assistant relies on.
If you already have access rules, Myriade reuses them.
Otherwise, Myriade reviews your data and suggests who should see what. Your data team validates. Then the rules apply in every assistant and every tool.
| Column | Finance | Marketing |
|---|---|---|
employee_id | Read | Masked |
department | Read | Read |
salary | Read | Deny |
bonus | Read | Deny |
Show me salaries by department.
Myriade checks your data continuously and alerts your data team when something breaks, and its agent traces the root cause before you open a ticket.
The result: errors get fixed before they reach an answer.
37 rows missing
13:00 load, alert fired at 13:05
Root cause, found by Myriade
The upstream job stg_orders failed at 12:58: column order_ts was renamed in the source.
Suggested fix
Remap the column, then rerun the load.
Without Myriade, an assistant spends four or five queries exploring tables before it can answer; with Myriade, it reads the documented context once and answers in one.
The result: lower cost, faster answers.
Every answer comes from a documented table, with its source.
Finance, marketing and stores get the same numbers.
Each team sees what it is allowed to see, in every tool.
In two months, we had an intelligent analytics engine running on our data.
Moez Hamad Chief Data & AI Officer, Jules
I was blown away immediately… Myriade lets me do things I couldn’t otherwise.
Allan Lemoine Head of Logistics, Jules
Before, for every analysis, I had to ask our developer. Now, I can do more myself.
Sandrine Eklou COO, Ekip
Nearly half of what teams ask are business questions. Each one used to be a request to the data team, or one more dashboard.
Business users58.7%
Data teams41.3%
Your warehouse’s built-in AI only knows that warehouse. Myriade works across all your sources, with every assistant your teams use.
“Myriade is hosted on our side, so it’s more secure. It’s connected directly to our production databases, and we’re comfortable using it.”