Put your data in
and your AI can use it
Managed PostgreSQL with MCP, REST and vector search built in Contracts, invoicing and support handled in Japan
One plan at ¥6,600/mo (tax incl.) · start the same day · commercial use OK
$ cat .mcp.json
"nwdb": { "url": "https://api.nwdb.dev/api/v1/mcp" }
✓ 11 tools registered
> Show me users likely to churn, by similarity
→ search_vectors running 0.3s
✓ Top 24 rows returned. Suggesting retention actions.
PAINS
Four walls every team hits when building an AI data layer
Data is scattered, so AI can't reason across it
It sits in a database, spreadsheets and SaaS tools, never in a shape you can hand to a model
You want Claude and Cursor to reach the data directly
Wiring an agent up to a database is heavy work to build yourself
Security work eats the schedule
2FA, WAF, backups, rate limits — the defence burns the time meant for the product
GDPR and Japan's APPI are a headache
Access control and an audit trail are too late if you bolt them on afterwards
SOLUTIONS
An answer for each wall
What you store is ready for AI and analytics as-is
- ▸MCP native — drive it straight from Claude or Cursor
- ▸Similarity search with pgvector, included
- ▸Funnel, RFM, LTV and retention in one request
- ▸Daily revenue and other rollups without a warehouse
Zero security code, production quality
- ▸TOTP two-factor authentication
- ▸Cloud Armor with Adaptive Protection
- ▸Daily backups held in Japan
- ▸SQL injection defence and rate limiting by default
RBAC, audit logs and tenant isolation
- ▸Every API call written to the audit log
- ▸Tenant isolation per workspace
- ▸SSO, VPC and CMEK available on Enterprise
- ▸Data held in Japan
ARCHITECTURE
One backend, every entry point
Store it once and reach it the same way from anywhere
GATEWAY
MCP, REST and SQL gateway
AI inference, vector search and live analytics
DB LAYER
PostgreSQL + pgvector + file storage
Daily backups held in Japan
BENEFITS
What your team gets back
5 min
to your first query
0 lines
of security code
None
no separate warehouse
Default
audit log on every API
Japan
contract, billing and support
DEMO
From connected to analysed in 90 seconds
-- Funnel over the last 30 days
SELECT step, count(distinct user_id)
FROM events GROUP BY step;
✓ 0.4s · 4 rows
QUICKSTART
Three steps to an MCP connection
- 1
Create a workspace
30 seconds after signup
- 2
Issue an API key
nwdb_sk_••••••••
- 3
Register it in Claude Code or Cursor
Add the MCP URL to .mcp.json
PRICING
One plan, nothing to choose
The same terms for individuals and companies Self-serve from signup to first query, with no sales cycle
STANDARD
- 6 GB database
- 6 GB file storage
- 1.5M API calls / month
- 20 GB egress / month
- 100K vector items
Every security feature is included
Sign up online and start immediately Billed monthly in advance, cancel any time Usage past the plan limits is metered
ENTERPRISE
- ▸SSO, VPC and CMEK
- ▸Reseller programme with published base rates and caps
- ▸Named support and an SLA
- ▸Invoice billing
DISCOVERY CALL
Start with a 30-minute call
Requirements, security and reseller terms — answered on the call