Jestor

Jestor is a Brazilian no-code platform for internal tools built on relational tables. Agents list, fetch, create, update and delete records, read table schemas and field options, manage users and tasks, and upload files.

Verified 15 actions API key available

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

A nightly reconciliation runs against a Jestor table of vendor records

Get Table Schema returns the generated field keys, because a field labeled Created by can be stored under a key nobody would guess

Get Field Options resolves each human readable choice to the option id a select field write actually stores

Count Records sizes the batch before the agent commits to paging through it

Update Record writes only the fields that changed and leaves everything else on the row untouched

The agent summarizes what it reconciled and names the rows it deliberately did not touch

An operator reviews any sweep whose count came back at Jestor's reporting ceiling, because at that point the true size of the table is unknown

What This Integration Enables

Jestor is a Brazilian no-code platform for internal tools, and the thing to understand about it before writing a single record is that the names on screen are not the names in the API. A field labeled Created by can be stored under a Portuguese key, a record in a vendors table is identified by a column named after the table rather than by a plain id, and a select field stores an option id rather than the text of the choice. None of these fail loudly. FlowRunner agents therefore start where the platform expects them to: Get Table Schema for the field keys and relationships, Get Field Options for the choice lists, then the record operations.

From there the surface is straightforward and genuinely useful. Agents list and filter records with AND and OR grouping, fetch nested related records inline rather than making a call per relationship, count a table before paging it, create workspace users, read the task list, and move files in and out of attachment fields including files the flow generated itself. What they do not do is decide on their own that a destructive sweep is safe, because one of Jestor's own protective behaviors makes that judgment unreliable. That is where human-in-the-loop enters, and it enters for a concrete reason rather than as a policy.

Without FlowRunner

Labels mistaken for keys A write aimed at the column name people see returns success and changes nothing on the record

Select fields written as text A choice field takes an option id, so writing the visible label leaves the cell empty or wrong

Internal tools drift from the systems they mirror The Jestor table and the source of record diverge, and nobody notices until a report disagrees

With FlowRunner

Schema read before every write Get Table Schema supplies the current keys, types and relationships at run time

Choices resolved, not guessed Get Field Options turns a label into the id Jestor stores, so select fields land correctly

Reconciliation as a running process Agents compare, write the delta, and escalate the rows they cannot resolve

Use Case Scenarios

The internal tool that stays in step with the system of record

An operations team runs vendor management in Jestor while the ledger lives elsewhere. On a schedule the agent reads the current state from PostgreSQL, calls Get Table Schema so the mapping follows today's field keys, and uses List Records with a filter to find the Jestor rows in scope. Where a value differs it calls Update Record for that field alone. Where a vendor is missing entirely it calls Create Record and captures the new id from the response. A digest posts to Slack with the counts and any row it skipped.
Documents the flow generated, attached to the record that needs them

A workflow renders a signed agreement returned from DocuSign and needs it on the Jestor record. Because the content was produced inside the flow rather than fetched from a public URL, the agent calls Create File From Base64 to turn it into a Jestor file object, then Update Record to write the reference into the attachment field. Later, when a person asks for the document, Download File pulls it back into FlowRunner file storage so a subsequent step can email it or push it to an archive.
Onboarding into the workspace without guessing at access

A new hire needs a Jestor login. The agent reads an existing user on the same team with List Users to learn which profile that role actually carries, rather than inventing one, then calls Create User with the resolved profile. The profile decides what the person can see and do, so the agent presents the profile it read and the user it is about to create before the call runs, and a workspace owner confirms.

Human-in-Loop Highlight

Jestor protects its own database by capping Count Records at 100,000. That is a sensible vendor decision and a trap for an autonomous agent, because a large table reports the ceiling rather than its real size. An agent that reads the count, concludes the table is manageable, and starts a delete sweep is reasoning off a number that means "at least this many," and Delete Record has no undo on this API. So the sweep agent treats the ceiling as a stop condition rather than a size. It posts: "The reconciliation would delete records matching an inactive status in the vendors table. Count Records returned exactly 100000, which is Jestor's reporting cap, so the number of rows this filter actually matches is unknown. Narrow the filter, run it in bounded pages with a review after each, or cancel?" A person answers. This is the platform thesis in miniature: the agent is not stopping because deleting is scary, it is stopping because it can prove it does not know how much it would delete.

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

15 actions

Records

List Records Returns records from a table with filtering, sorting, paging and field selection. Filters join with AND by default, and a group can branch to OR. Naming only the fields you need improves performance materially on wide tables.
Get Record Returns one record by its numeric id. Turning the summarized option off pulls in the related child records as well, which is the difference between a lookup and a full object graph.
Create Record Creates a record from a map of field API names to values and returns it with its new id, under a key named after the table.
Update Record Updates a record, writing only the fields supplied. Jestor identifies records by a column named after the table, and this operation builds that key from the table and record id so a workflow never has to embed it.
Delete Record Permanently deletes a record by its numeric id. There is no undo on this API.
Fetch Records Fetches records with the option to pull nested related records in the same request, which the plain listing cannot do. System tables such as users and tasks are excluded from it.
Count Records Returns how many records a table holds, optionally matching a filter. The vendor caps the count to protect the database, so a table above the cap reports the cap rather than its true size.

Schema

Get Table Schema Returns a table's structure: every field with its display label, API key, type, whether it is required, and its relationships. The first call in any workflow that writes records.
Get Field Options Returns the configured options behind tag, radio select and multiple select fields, each with its id and label. Used to resolve a human readable choice to the id Jestor stores.

Users and Tasks

List Users Returns the users in the workspace with sensitive fields stripped. Users are a system table and are not reachable through the record operations, so this is the only way to read them.
Create User Creates a workspace user with an email, a password and a profile. The profile decides what the user can see and do, so read it from an existing user rather than guessing.
List Tasks Returns the tasks in the workspace. Like users, tasks are a system table outside the record operations.

Files

Upload File Uploads a file into a table from a URL and returns a file reference. The upload alone does not attach the file to a record.
Create File From Base64 Creates a file object directly from content generated inside the flow, such as a rendered PDF or an export, and returns the reference to write into an attachment field.
Download File Downloads a file from an attachment field into FlowRunner file storage and returns a URL a later step can use.

Frequently Asked Questions

What can FlowRunner do with Jestor?

FlowRunner agents can run List Records, Get Record, and Create Record in Jestor, plus 12 more actions.

Does connecting Jestor to FlowRunner require OAuth?

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

Can Jestor trigger a FlowRunner workflow automatically?

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

Start building with Jestor

Free plan, no card required. Connect in minutes.

Start Building Free Talk to us