Pre-beta · free while we build

Analytical memory for AI agents.

A real database your agent writes to and you query in plain English — or connect the ClickHouse you already run. Aggregate over everything it remembered, not just the top-k a vector store recalls.

Connect your warehouse → See both ways in

No account needed to try — just describe your data →

› how many orders today, by drink

91 orders today across three drinks.

Drink	Orders
Latte	42
Cappuccino	31
Cold brew	18

cache· plan· validate· exec· summarize 240 ms

A representative reply — every answer comes back as plain English, the data, and a trace you can open.

Two ways to start

Pick the door that fits.

Use it as agent memory

Give your agent a real Postgres database to write to as it learns — typed rows, queried in English. Install it in your host over MCP.

Install in your agent.

Add to Cursor
Opens Cursor and asks you to confirm before writing the config.
If Cursor didn't open, paste this config manually:
Add to VS Code
Opens VS Code and asks you to confirm before writing the config.
If VS Code didn't open, paste this config manually:
Run in your terminal.
Run in your terminal.
```
claude mcp add --transport http nlqdb https://mcp.nlqdb.com/mcp
```
Add to ~/.codex/config.toml.
Add to ~/.codex/config.toml.
```
[mcp_servers.nlqdb]
url = "https://mcp.nlqdb.com/mcp"
```
Settings → Connectors → Add custom connector — paste the URL.
Settings → Connectors → Add custom connector — paste the URL.
```
{
  "mcpServers": {
    "nlqdb": {
      "url": "https://mcp.nlqdb.com/mcp"
    }
  }
}
```
Cascade → MCPs → Configure, or ~/.codeium/windsurf/mcp_config.json.
Cascade → MCPs → Configure, or ~/.codeium/windsurf/mcp_config.json.
```
{
  "mcpServers": {
    "nlqdb": {
      "serverUrl": "https://mcp.nlqdb.com/mcp"
    }
  }
}
```
Agent Panel → Add Custom Server, or ~/.config/zed/settings.json.
Agent Panel → Add Custom Server, or ~/.config/zed/settings.json.
```
{
  "context_servers": {
    "nlqdb": {
      "url": "https://mcp.nlqdb.com/mcp"
    }
  }
}
```

Anonymous — the configs ship with the pk_live_REPLACE_ME placeholder. Sign in (free) to inline your live key →

Not building an agent? just describe your data →

Question your ClickHouse

Point nlqdb at the warehouse you already run — ClickHouse or Postgres. It reads your schema; you ask in English.

connection string clickhouse://reader:••••••••@warehouse:9440/analytics sealed

Your connection string is sealed server-side — never stored in your browser.

Connect your ClickHouse →

No login wall

Describe your data and get a working answer before you ever sign in.
Read-only by default

Connect with a read-only string; writes are previewed before they run.
Every answer shows its work

A trace with timings sits under each reply — honest latency, no spinner-lies.

A 3D map of the nlqdb engine: three lit zones — you ask, the engine, your data — connected by flowing jade request paths.

Under the hood

One engine, seen from above.

Fly through the request path — five surfaces, one engine, your data — and click every moving piece. Explore the 3D map →

Analytical memory for AI agents.

Use it as agent memory

Question your ClickHouse

One engine, seen from above.

Something broke on this page.