Operations & Internal Tools
Connect Retool and Supabase
Screens in Retool, data in a Postgres database you own and can expose safely.
- Data flows
- 4 automated, two-way where safe
- Built with
- Retool, Supabase Postgres, Supabase Row Level Security
- Timeline
- Live in 3 to 5 business days from scope
- Owned by
- You. Accounts, credentials and code in your name.
In short
Connecting Retool to Supabase puts your operational data in a Postgres database you own, with Retool as the screens on top. SBD uses Supabase to publish only what outside services need, as in a WhatsApp catalogue build, and as the landing point when a system outgrows Retool and moves to custom code.
The workflows
What runs between Retool and Supabase.
Each line is one automated workflow. An event in one tool, the checks and mapping in between, the action in the other, and the person who hears about it.
-
Public catalogue without exposing internal data
WhenThe team edits a product in the Retool app
Automation- Verify and validate
- Map and normalise
- Upsert by key
ThenCopy only the public fields to a Supabase catalogue table
And- Audit log entry
- Alert if it fails or goes quiet
- Review queue if unsure
-
WhatsApp catalogue kept current
WhenThe Supabase catalogue table changes
Automation- Verify and validate
- Map and normalise
- Upsert by key
ThenSync the Meta commerce catalogue from Supabase, never from the Retool database
And- Audit log entry
- Alert if it fails or goes quiet
- Review queue if unsure
-
Copy the Retool data before moving
WhenA move off Retool is approved
Automation- Verify and validate
- Map and normalise
- Upsert by key
ThenRun a one-way sync that copies the Retool database into Supabase on demand, while the old app keeps working
And- Audit log entry
- Alert if it fails or goes quiet
- Review queue if unsure
-
A cut-over with the old app kept read-only
WhenCut-over day
Automation- Verify and validate
- Map and normalise
- Upsert by key
ThenFinal sync, the Retool app set to read-only, and users switched to the new app in one step
And- Audit log entry
- Alert if it fails or goes quiet
- Review queue if unsure
Every record passes through the same stages
- 1Receive
Webhook or polling from the source tool, with a signed payload and a replay log.
- 2Validate
Required fields present, types right, record key found. Anything else goes to the review queue.
- 3Map
Fields translated to the destination schema. Enumerations, currencies and dates normalised.
- 4Upsert
Written against the record key, so a retry updates rather than duplicates.
- 5Confirm
Outcome logged. Alert if the count goes quiet. Reconciliation run nightly.
Read this first
Why the obvious route breaks.
A one-step Zap connects Retool and Supabase in an afternoon. This is what goes wrong once real volume and real edge cases arrive.
The direct connection string does not reach Retool
Supabase direct connections map to an IPv6 address, and Supabase lists Retool among services that only accept IPv4. The shared pooler is IPv4, so the connection has to go through it, in the right mode, or the resource never connects.
Poolers and prepared statements collide
Retool turns Postgres queries into prepared statements by default. Retool advises disabling that only for setups such as a transaction-mode pooler, and doing so removes its SQL injection protection. Retool also warns that a pooler can break access policy enforcement. The mode is a design decision, not a checkbox.
The secret key ignores your row rules
A Supabase secret key, or the legacy service_role key, bypasses Row Level Security. Pasting it into a Retool resource gives every app on that resource full access. Publishable keys follow RLS, which Supabase says should be on for every table before you deploy.
Outside services should never see the internal tables
In the WhatsApp product promotion build, the Meta catalogue reads from a Supabase database so the internal Retool database is never exposed. Pointing an outside service at the working tables leaks fields that were never meant to leave.
What it costs today
The gap, as the team feels it and as the engineer sees it.
In the team
- Outside services need data the team does not want exposed
- Business rules live in queries nobody can review
- A workflow stopped and nobody was told
- Per-user cost climbs as more people need access
- Fear of losing data blocks any move
Under the hood
- IPv6 direct connections versus IPv4-only Retool
- Prepared statements and pooler modes
- Secret keys bypass Row Level Security
- Default 1,000-row cap on Data API reads
- Keys written into queries in plain text
What we wire up
The data flows.
- 01 When
The team edits a product in the Retool app
ThenCopy only the public fields to a Supabase catalogue table
- 02 When
The Supabase catalogue table changes
ThenSync the Meta commerce catalogue from Supabase, never from the Retool database
- 03 When
A move off Retool is approved
ThenRun a one-way sync that copies the Retool database into Supabase on demand, while the old app keeps working
- 04 When
Cut-over day
ThenFinal sync, the Retool app set to read-only, and users switched to the new app in one step
Used for
- Public catalogue without exposing internal data
- WhatsApp catalogue kept current
- Copy the Retool data before moving
- A cut-over with the old app kept read-only
- Retool on a database you own
- Supabase as a Retool resource
- Moving off Retool to custom code
- Row Level Security for outside access
- Audit log on every table
- Demo database seeded for reviewers
Built with
Data map
What becomes what, and what it is keyed on.
| In Retool | In Supabase | Matched on | Direction |
|---|---|---|---|
| Products, categories and prices in Retool | Supabase catalogue table, public fields only | Product ID | one-way |
| Recipient lists in Retool | Kept internal, not published | List ID | one-way |
| Retool database tables | Supabase schema with keys and RLS | Primary key per table | one-way |
| Retool app actions after cut-over | Supabase tables with an audit log | Row ID | two-way |
Field-level mapping is agreed in the blueprint and written into the runbook. Two-way rows carry loop protection so a write from the sync never re-triggers the sync.
In every build
What you get with the sync, whichever pair.
Under the hood
- Upserts keyed on the record
Every write is matched on the source ID. A retry or a re-fire updates the existing record instead of creating a second one.
- Retries with backoff
Rate limits and timeouts are retried with increasing delays, then parked for review, never dropped.
- Idempotent, loop-safe writes
Two-way flows carry a marker so a change written by the sync does not re-trigger the sync.
- Reconciliation
A nightly job compares counts and totals on both sides and lists any record that differs.
- Audit log
Every run records input, mapping result, action taken and who or what triggered it. Searchable by record.
- Error monitoring
Loud failures alert within minutes. Quiet ones are caught by an outcome count that alerts when it stops moving.
- Review queue
Records the sync is not sure about land in a screen a person can read, correct and release.
- Scoped credentials
One credential per connection, limited to the objects it needs, stored encrypted, in your account.
For the team
- Slack or email notifications
The alerts go to a named person, not a channel nobody reads. You choose Slack, email, WhatsApp or all three.
- A runbook
One document per integration: what it does, what each alert means, and what to do in the first ten minutes.
- Training
A live session with the people who will use the review queue, recorded for whoever joins later.
- Handover
Admin access, billing and code sit in accounts in your name from day one. If we leave, it keeps running.
- Sixty days of hypercare
The engineers who built it stay on call after go-live, fix what breaks and tune what the team asks for.
- A monthly report
Runs, records synced, exceptions handled, and what changed. Plain language, one page.
Security and your data
Where the data goes, and who holds the keys.
- Accounts
- Every platform account is created in your name. We work through managed access you can revoke.
- Credentials
- Stored encrypted in the orchestration layer. Scoped per connection. Rotated on handover.
- Data in transit
- TLS everywhere. Webhooks verified by signature before anything is processed.
- Data at rest
- Only what the sync needs. No copies of your customer data on our machines.
- AI steps
- Where an integration uses a model, it is through a business API that does not train on your data.
- Retention
- Run logs kept 90 days by default, adjustable. Payloads redacted of card and ID numbers.
- Agreements
- DPA and NDA on request. Regional rules covered in our privacy policy.
Keys and sensitive data
How API keys, tokens and personal data are handled.
- Keys live in a vault, never in code
API keys and tokens are stored in the orchestration layer's encrypted credential store or the platform's secret manager. None are written into workflows, repos, spreadsheets or chat.
- One credential per connection
Each integration gets its own key, scoped to the objects and actions it needs. A lead sync cannot read invoices. Revoking one key stops one thing.
- Service accounts, not personal logins
Connections use a dedicated service user in your account, so nothing breaks when a person leaves and nothing runs under someone's name.
- Two-factor on every account
Every platform account we set up has two-factor authentication on, with recovery codes handed to you, not kept by us.
- Rotation on handover and on request
Keys are rotated when the build is handed over, when anyone with access leaves, and at any time you ask.
- Secrets never reach the logs
Audit logs redact tokens, card numbers, government IDs and anything you flag as sensitive before a line is written.
- Webhooks verified before processing
Inbound calls are checked against the sending tool's signature. Unsigned or stale payloads are dropped and counted.
- Access you can see and revoke
Our access is through managed roles in your accounts. You can list it, limit it, and remove it in one click without touching the build.
Engineering decisions
Decided before the first workflow is written.
- Rate limits and backoff
Every API has a ceiling. Calls are batched, queued and retried with increasing delays so a busy hour does not become a failed one.
- Backfill and pagination
The first run has to load history, not just new events. Backfills page through the source in order, resume after interruption, and never double-write.
- Out-of-order and duplicate events
Webhooks arrive twice and out of sequence. Each event carries a version or timestamp, and older versions never overwrite newer state.
- Schema changes
A renamed field or a new dropdown value is the most common silent failure. Payloads are validated against a contract, and a mismatch stops the run and raises a flag.
- Time zones, currencies, units
Dates are stored in UTC and shown local. Money carries its currency. Quantities carry their unit. Conversions happen once, in the map step.
- Sandbox first
Builds run against sandbox or test accounts with seeded data before they touch production. Go-live is a switch, not a rewrite.
- Versioned and reversible
Workflows and code are version-controlled. A release can be rolled back in minutes without losing the records processed since.
- API deprecations watched
Tools retire API versions on a schedule. We track the dates for every connection in the runbook and upgrade before the cut-off.
- Cost per run known
Each workflow has a cost per execution, including any model calls, logged alongside the run. You see the bill before it arrives.
- Metrics, not just errors
Runs per hour, records in and out, exception rate and latency are charted. Trends show a problem before an alert does.
Proof
Where we have done this.
The offer
Hire SBD to connect Retool and Supabase.
A fixed scope, agreed in a discovery call, and live in three to five business days. Accounts in your name. The engineers who built it on call for sixty days after.
- ScopeOne call to agree the flows, the fields and the owner. Written up the same day.
- Build4 workflows, the data map above, validation, upserts, retries and reconciliation.
- Go-liveSandbox first, then a switch. Day three to five from scope for most pairs; longer only if a tool needs approval, such as WhatsApp templates.
- HandoverRunbook, training session, alerts to a named owner, admin access and code in your accounts.
- AfterSixty days of hypercare, then a care plan or run it yourself.
Questions
What people ask about this pair.
Can we keep using Retool Database instead?
Yes. Retool Database is Postgres, and Retool gives admins a connection string to reach it from outside. We still put anything outside services read into a separate Supabase database, so the working data and the published data never share credentials.
Why Supabase rather than any Postgres host?
Any Postgres works for Retool. Supabase adds Row Level Security tied to its keys, an API for outside services, scheduled jobs and storage, which is what the catalogue and migration builds needed. The data stays in standard Postgres either way.
Does moving to Supabase mean leaving Retool?
No. Plenty of teams should stay on Retool. Putting the data in Supabase first means a later move is a rebuild of screens rather than a data rescue, which is what made our three moves off Retool in 2026 practical. Internal tools start from $3,000.
Do Supabase reads have a row limit?
The Supabase Data API returns at most 1,000 rows per request by default, and the limit is set in project settings. Anything that reads a whole table, such as a catalogue sync, pages through it rather than trusting one read.
Related
-
Airtable and Retool
Airtable is a good database and a limited interface. Retool fixes the second half.
-
Supabase and Resend
Email that leaves your app once, from your own domain, with proof it arrived.
-
WhatsApp Business API and Zoho Books
The invoice raised where the accountant expects it, and everyone who moves the order told on WhatsApp.
- Solution
IT
Every workflow we automate for this kind of team.
Retool and Supabase
Keep Retool for the screens. Own the database underneath.
Retool connected to Supabase the right way, safe publishing to outside services, and a path off Retool when the system outgrows it, built in three to five business days.
Start with a discovery call