Tabidoo

Tabidoo is a Czech low-code database platform organized as applications containing tables containing records. Agents list applications and tables, search, count and summarize records, create, update, replace and delete records in bulk and check the current user.

Verified 23 actions API key available

Tabidoo website Platform Documentation Capability data verified 2026-08-27

A nightly synchronization run starts against the customers table

Get Table returns the internal field names, which are not the column labels shown in the interface

Count Records reports how many records the intended filter actually matches

The agent compares that count against the size of the incoming batch

Update Records Bulk writes the changed fields in one request, leaving unlisted fields alone

A summary of created, updated, and skipped records posts to the operations channel

The data owner reviews any run where the matched count and the batch size disagree

What This Integration Enables

Tabidoo is a Czech low-code database organized as applications containing tables containing records, and it is the one in this group with an opinion about how change should happen. Create Development App produces a development copy of a production application, which is the vendor's supported way to alter a schema without touching live data, and the recommended path runs development, then test, then production. Most no-code databases treat that as the customer's problem. Building automation on a platform that treats it as a first-class operation is materially different, because a flow can provision the copy, exercise it, and report on it before anything reaches the application people are using.

The rest of the API rewards the same care. Bulk operations exist for create, update, replace, and delete, and the vendor recommends them over looping the single ones. Filters are a string with the operator in parentheses, which this connector assembles from structured conditions so a hand-built expression cannot be subtly wrong, because a malformed Tabidoo filter is ignored rather than refused and quietly returns the whole table. Count Records and Search Record Count exist precisely so a flow can size a change before making it, and FlowRunner agents use them that way: read, count, show the number, then write. That sequence is what human-in-the-loop looks like on a database connector, and it is the default rather than something you add after an incident.

Without FlowRunner

Filters that fail open A malformed filter is ignored rather than refused, so a wrong condition returns the whole table and reads as a match

Labels mistaken for fields Writes are mapped against the column headings on screen, and a field whose internal name differs is silently never written

Schema changes on live data Adding or reshaping a table means changing the one everyone is using and hoping nothing depended on the old shape

With FlowRunner

Counted before written Count Records runs the same filter first, so the size of a change is known before the change happens

Internal names resolved Get Table reports both the label and the internal name, and the write uses the one the API accepts

A place to change things Create Development App produces a copy to change, which is the vendor's own path from development to test to production

Use Case Scenarios

A synchronization that knows its own size

Overnight, a customer list arrives from HubSpot. The agent calls Get Table to resolve internal field names, then Count Records with the filter that identifies the records this run should touch. If that count is close to the incoming batch, Update Records Bulk applies the changes in a single request and the flow completes unattended. If the count is wildly larger, that is the signature of a filter that failed open, and the run stops before writing anything. The check costs one request and is the difference between a synchronization and an overwrite.
Reporting without pulling the table

A weekly operations report needs totals by category. Rather than paging every record into the flow to add a column up, the agent calls Get Data Summary, which computes the aggregate on Tabidoo's side with the grouping the report needs. The figures land in a sheet in Google Sheets and a short digest posts to Slack. Search Record Count covers the cases where the filter is long enough that a query string would not hold it.
Changing the schema without changing the live one

A team needs a new field and a reshaped relationship. The agent calls Create Development App to produce a copy, applies the change there, and runs the flow's own reads and writes against the copy to confirm they still work. Get Application TypeScript Definition returns the most compact machine-readable description of the resulting schema, which the flow diffs against the production one and posts for review. Nothing touches the live application until a person has read that diff.

Human-in-Loop Highlight

Tabidoo's most destructive write is not its delete. Delete Records Bulk already carries a guard: this connector requires both a non-empty filter and a Limit, so an unbound template expression cannot become a whole-table clear and a wrong filter costs at most the capped number of rows. Replace Records Bulk carries no such requirement, and it is worse in a way that is easier to miss. For each entry in the batch, the fields you send become the record and every field you leave out is cleared. It reports success. The rows are still there, still countable, and quietly emptied of everything the payload did not mention.

So that is where the agent stops. It runs Count Records with the same filter, reads Get Table for the full field list, and asks: "Replace Records Bulk on 1,206 matched records will write the 6 fields in this payload and clear the other 19 on every one of them, including the internal notes and the assigned owner. Update Records Bulk would write the 6 and leave the rest. Which did you mean?" A person picks the verb. Almost every time they meant the partial one, which is exactly why the question is worth asking: the operation that looks like an ordinary batch update is the one with no undo behind it.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via preferred channel

Human decides

Agent resumes with decision

Agent Capabilities

23 actions

Schema

List Applications Returns the applications the token can reach, each with the identifier every other operation needs. A token carries exactly its owner's privileges, so this lists what that user can see rather than everything on the account.
Get Application Returns one application with its tables and their definitions. Used to discover table identifiers before reading or writing records.
List Tables Returns the tables in an application, each with its identifier and name.
Get Table Returns a table's full definition: every field with its internal name, type, and settings, plus attached scripts. Called before writing, because internal field names differ from the displayed labels and both filters and record bodies use the internal one.
Get Application TypeScript Definition Returns a TypeScript definition of an application's schema. The most compact machine-readable description of every table and field, which makes it useful for generating and diffing mappings.
Create Development App Creates a development copy of a production application. This is the vendor's supported way to change a schema without touching live data, on the recommended path from development to test to production.

Records

List Records Returns records with filtering, sorting, and paging. Conditions may be given as structured objects and are assembled into Tabidoo's own expression syntax, which is where a hand-written filter usually goes wrong. Limit accepts up to 1,000 and defaults to 25.
Search Records Returns records using a filter sent in the request body rather than in the URL. The route to use when a filter is long or carries many values.
Count Records Returns how many records match a filter, without returning them. Much cheaper than fetching records to count them, and the right way to check a filter before running it.
Search Record Count Returns how many records match a filter sent in the body, for the same reason Search Records exists.
Get Data Summary Computes aggregates over a table on Tabidoo's side, grouped as specified, rather than pulling every record into the flow to add a column up.
Get Record Returns one record by identifier with all of its field values.
Create Record Creates one record from a map of internal field names to values. The response type controls how much comes back, which is worth setting to the identifier alone on a high volume flow.
Update Record Updates only the fields supplied and leaves the rest untouched. The safe edit and the default.
Replace Record Replaces a record wholesale: the fields sent become the record and every field left out is cleared. Reach for it only when the record is genuinely meant to be exactly what is being sent.
Delete Record Permanently deletes one record. There is no undo on this API.

Bulk Records

Create Records Bulk Creates many records in a single request. The vendor recommends the bulk routes over looping the single ones: fewer requests, lower latency, and much less chance of hitting the per-minute rate limit.
Update Records Bulk Updates many records in one request, writing only the fields each entry names and leaving the rest untouched. Each entry carries its record identifier alongside its fields.
Replace Records Bulk Replaces many records wholesale in one request, clearing every field each entry leaves out. The most destructive write on this API short of a delete, and the operation this page's human gate exists for.
Delete Records Bulk Deletes every record matching a filter, up to a Limit. Both are required on purpose: an empty filter would match the entire table, and the Limit caps what a mistaken filter can remove.

Files

Get Image Thumbnail Downloads the thumbnail of an image stored in a record's file field. Used to show a preview without transferring the full sized original.

Account

Get Current User Returns the user the token belongs to. The first thing to check when an operation is refused, because the token may be valid and simply belong to someone who cannot see the application.
Ping Verifies that the token reaches Tabidoo. It takes no input and touches no data, which makes it the right first step when separating a credential problem from a data one.

Frequently Asked Questions

What can FlowRunner do with Tabidoo?

FlowRunner agents can run List Applications, Get Application, and List Tables in Tabidoo, plus 20 more actions.

Does connecting Tabidoo to FlowRunner require OAuth?

No. Tabidoo connects to FlowRunner with an API key, no OAuth flow required.

Can Tabidoo trigger a FlowRunner workflow automatically?

Tabidoo doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with Tabidoo

Free plan, no card required. Connect in minutes.

Start Building Free Talk to us