Supalens
Log inConnect Supabase
THE AI DATA OPERATOR FOR SUPABASE TEAMS

Your business has signals.Know which ones matter.

Monitor the metrics behind your product. Investigate what changed. See the evidence. Get back to building.

Connect SupabaseSee how it works

Read-only data access · Human-approved monitors

Connects to
Sample workspace/Overview

Business overview

A clear view of what needs your attention.

One signal needs your attention

3 completed payments are missing an active entitlement.

Active monitors
3
Healthy
2
Needs attention
1

Priority monitors

Paid without access
3 ordersIssue
Paid job failure rate
0.4%Healthy
Trial conversion
8.2%Healthy

Illustrative values. No production database connected.

Investigation

Payments succeeded. Access didn’t.

3 orders have no matching active entitlement. Start with the evidence.

Confirmed in data

Payment status: completed

Active entitlement: not found

Next check

Unverified hypothesis: provisioning may be delayed or missing. Review worker logs and retries; these records alone do not establish the root cause.

Product preview · sample dataTry the live demo

Infrastructure can be healthy. Your customer experience might not be.

  • STACKSupabase + GitHub + Slack

    The database says what exists, the code says what to count, Slack says where to tell you.

  • READ-ONLYRead-only by design

    There is no tool that writes. SELECT only, inside a READ ONLY transaction.

  • VERIFIEDKPIs verified before monitoring

    Every proposed KPI comes from SQL that actually ran.

Connects to

Bring the database you already run

Connect in minutes with read-only access. Supalens never needs a write permission, and your credentials stay encrypted.

  • Supabase

    One-click OAuth. A read-only role is created for each project.

    • Plain-language questionsSupported
    • KPI proposalsSupported
    • Monitors and alertsSupported

  • PostgreSQL

    Connection details for a read-only user. Self-hosted, Neon or Amazon RDS.

    • Plain-language questionsSupported
    • KPI proposalsSupported
    • Monitors and alertsSupported

  • MySQL

    Connection details for a read-only user. Self-hosted, PlanetScale or Amazon RDS.

    • Plain-language questionsSupported
    • KPI proposalsSupported
    • Monitors and alertsSupported

  • Firestore

    A Google service account with Datastore viewer access.

    • Plain-language questionsSupported
    • KPI proposalsNot supported
    • Monitors and alertsNot supported

GitHub code context can be added to any source. KPI proposals and monitors need a SQL database.

FROM BUSINESS CONTEXT TO CLEAR NEXT STEPS

Less digging.More understanding.

For founders and small teams running on Supabase. Follow the business signals your infrastructure metrics don’t explain.

  • Paid without access
  • Failed paid jobs
  • Changes in conversion
01

Start with your context.

Connect your database and repository. Review what Supalens understood so proposed monitors reflect how your business works.

02

Choose what matters.

Review the definition and threshold before a monitor runs. Keep the business logic visible and approve what your team needs.

03

Go from signal to evidence.

Review what was observed, what is still uncertain, and where to look next. Bring the signal into Slack or email.

Monitoring · Root cause

The signal, in your workflow.

When an approved monitor fires, Supalens investigates the data and stored code context. Review confirmed facts and likely causes separately in Slack or email. An all-clear follows when the metric recovers.

Try the live demo

Alert firing2:14 PM

Daily revenue below baseline

₩3.1M−26%Baseline ₩4.2M

Metric
daily_revenue
Condition
< ₩4.0M
Source
DailyCart · prod
Confirmed in data
  • Payment failure rate 0.3% → 7.8%
  • Renewal revenue −31% · new subscriptions −4%
  • Japan −52% · US normal
Likely driver

Failures at the payment-confirmation step most likely rose after the coupon migration shipped on 7/5 (0051_coupons.sql). Per the code context, review the checkout confirmation logic first.

Investigated by Supalens · the driver is a hypothesis to review

  1. Slack · #alerts2:14 PM

    Alert firing

    Confirmed in data · Likely driver

  2. Mail2:14 PM

    [supalens] Daily revenue anomaly detected

    owner@dailycart.co.kr

  3. On recovery6:40 PM

    Alert resolved

    The value moved out of the alert condition and is back to normal. Current value ₩4.3M.

Illustrative values. No production database connected.

Security

By design, Supalens can’t change your data.

Read-only isn’t a promise — it’s the architecture. There’s no tool that writes, your credentials are sealed, and the analysis AI runs isolated, without them.

ENCRYPTION

Sealed credentials

Your database credentials are envelope-encrypted with a per-organization key (AES-256-GCM). The master key that wraps it is never stored in the database, so a database dump alone can’t decrypt them. Plaintext never reaches your browser or the AI model.

  • AES-256-GCM
  • per-tenant DEK
  • AAD-bound
ACCESS

Structurally read-only

Questions and monitors run read-only. Non-SELECT statements are rejected, and a READ ONLY transaction, statement timeout, and row cap are always applied. Supabase OAuth connections get a dedicated read-only role (supalens_ro) minted per project; directly connected databases run on the credentials you provide, wrapped in the same SELECT-only guard and READ ONLY transaction.

  • SELECT-only
  • READ ONLY tx
  • supalens_ro · OAuth
RUNTIME

Isolated AI execution

The analysis AI runs in a sandbox with credentials stripped from its environment — no shell, file, or network tools. Database credentials never enter the model process, and only pre-approved tools can run.

  • env-stripped
  • no shell / net
  • deny-by-default
ISOLATION

Tenant isolation

Row-level security is enabled on every user table. One analysis is bound to one organization, chosen by the server — never by the question or the model, so it can’t reach another org’s data.

  • RLS
  • 1 run = 1 tenant

In the open

We don’t replicate your whole database. During analysis, query results are sent to Anthropic (Claude), and we store the answer, a sample of up to 200 rows, your schema structure, the context extracted from code (SUPALENS.md), approved KPI definitions, and monitor events with their investigation results. Every query is audit-logged and auto-deleted after a set period, and deleting a data source or your account revokes the stored credentials at once — they can no longer be used.

Compliance

  • Privacy policy with full subprocessor disclosure
  • Data-subject rights: access, deletion, and more
  • Payments handled by Paddle — no card data stored
Privacy policy
Pricing

Pick by how many KPIs you watch.

Start free, move up as you monitor more.

Free

$0

When you start watching metrics alone

Includes:

  • 3 monitors
  • Checked every 2 hours
  • Email · Slack · Discord alerts
  • 1 data source
  • 1 team member
  • 15 questions / month · fair use
Start free
Recommended for startups

Basic

$89/mo

For SaaS teams just getting going

Everything in Free, plus:

  • 10 monitors
  • Checked every 10 minutes
  • Email · Slack · Discord alerts
  • 3 data sources
  • 5 team members
  • 150 questions / month · fair use
Get started

Standard

$299/mo

For teams with more to watch

Everything in Basic, plus:

  • 50 monitors
  • Checked every 5 minutes
  • Email · Slack · Discord alerts
  • 10 data sources
  • 15 team members
  • 750 questions / month · fair use
Start now

Premium

$899/mo

For large-scale operations, no limits

Everything in Standard, plus:

  • Unlimited monitors
  • Checked every 5 minutes
  • Email · Slack · Discord alerts
  • Unlimited data sources
  • Unlimited team members
  • Unlimited questions · fair use
Get started

All prices are in USD. Start free — no credit card required. AI investigations run when a monitor fires.

Frequently asked questions

Early-stage SaaS and AI SaaS teams without a data team. When a founder, CTO, or product lead has no time to build dashboards but still needs revenue, conversion, and churn watched, Supalens takes the data-operator seat.

Most AI SQL tools only see the schema. Supalens reads business rules from your GitHub code — “cancelled subscriptions are excluded from MRR”, “internal accounts are excluded from counts” — defines KPIs on top of them, and then keeps monitoring the approved KPIs and investigates when one breaks, instead of stopping at an answer.

Supabase connects via OAuth and a read-only role is minted per project to read the schema. GitHub connects via an App install; the repo is cloned read-only, business rules are extracted and stored as SUPALENS.md. Then Supalens proposes execution-verified KPIs, and the ones you approve get monitors.

Yes. Every KPI is proposed only from SQL that actually ran, and only the ones you approve become dashboards and monitors. Definitions are shown in plain language, and the SQL is always one click away.

We never replicate your database. We store answers, samples of up to 200 rows, your schema structure, the context extracted from code (SUPALENS.md), approved KPI definitions, and monitor events with their investigation results. Credentials are envelope-encrypted with a per-tenant key and never appear in plaintext in code, logs, or the model.

No. There is no tool that writes. Every query and monitor is SELECT-only inside a READ ONLY transaction; INSERT, UPDATE, DELETE, and DROP are blocked.

The cause is a hypothesis. Alerts separate “confirmed in data” from “likely driver”, and the SQL that ran is always shown. Base important decisions on the confirmed side — and a wrong hypothesis becomes the lead for the next investigation.

By the number of monitors you keep, how often they are checked, and your data sources and team members. Free watches 3 monitors every 2 hours; paid plans watch more monitors every 10 or 5 minutes. Natural-language questions are only a fair-use limit.

Yes. Plug in the MCP server mcp.supalens.ai to ask using the approved KPI definitions as-is, break a cause down, and create new monitors. It works without the web console.

Supabase (OAuth), PostgreSQL, MySQL, and Firebase Firestore. Automatic KPI proposals currently support the SQL family (Supabase, PostgreSQL, MySQL); GitHub code context can be attached to any source.

You build the product. Supalens watches the business.

Connect SupabaseTry the live demo

No credit card · Read-only by design