Geckoboard
Analytics & DataPush live business metrics into Geckoboard KPI dashboards from your flows. Agents define dataset schemas, then keep dashboards current by replacing or appending rows as sales, revenue, or conversion numbers change in source systems.
What This Integration Enables
A wall dashboard makes a promise: the number up there is true right now. Geckoboard renders the number; something still has to keep the promise. This connector makes that FlowRunner's job. Agents define dataset schemas with typed fields, run full refreshes with Replace Data, maintain rolling trend datasets with Append Data and unique keys, and verify connectivity before a pipeline goes live. The opinionated part is what happens before the push. A dashboard is a broadcast medium: whatever reaches it is instantly believed by everyone walking past. So FlowRunner pipelines validate against the source, compare against history, and hold anomalies for human review, because the cost of a wrong number on the wall is measured in meetings, and the cost of a dashboard people stop believing is measured in every decision made on gut feel afterward. - Define KPI dataset schemas with Number, Money, Date, Percentage, and Duration fields - Run scheduled full refreshes from warehouses, spreadsheets, and apps - Append timestamped snapshots with Unique By dedup and Delete By roll-off - Sanity-check batches against prior values before they publish - Verify credentials before wiring a pipeline into a live board
Without FlowRunner
With FlowRunner
Use Case Scenarios
Revenue on the wall, minutes behind reality
On a schedule, the agent pulls current figures from [Stripe](/integrations/stripe), payment volume, new subscriptions, refunds, and calls Append Data to add a timestamped row to the revenue dataset, created with Unique By on the date field so a rerun updates rather than duplicates. Delete By lets old rows roll off at the retention limit, so the trend chart maintains itself. The sales floor watches actual revenue move, not a weekly recap of it.
The spreadsheet that became a dashboard
Operations tracks fulfillment metrics in [Google Sheets](/integrations/google-sheets). When rows update, the agent reads the sheet and calls Replace Data for a full refresh of the ops dataset. The spreadsheet stays the working surface the team edits; the dashboard becomes its always-current projection. Nobody transcribes numbers between the two again, and the version on the TV cannot drift from the version in the sheet.
Standing up a new KPI in one flow
The leadership team wants deal-stage conversion on the wall. The agent calls Create or Replace Dataset defining the schema, stage name as String, conversion as Percentage, updated date as Date, then backfills history from the CRM with Append Data in 500-row batches, and schedules the nightly refresh. Verify Authentication ran first, so the pipeline was proven before it touched a live board. From request to living chart is one flow definition, not a data-engineering ticket.
Human-in-Loop Highlight
Delete Dataset permanently destroys a dataset and its full history, and every dashboard widget built on it goes dark at once, usually on a screen executives look at daily. There is no undo, and the blast radius is not visible from the API: the dataset ID says nothing about which boards consume it. The same trap hides in Create or Replace Dataset, where reusing an existing ID replaces the schema and orphans the data. So FlowRunner flows treat dataset identity as protected: agents append and refresh freely, but any Delete Dataset call, and any Create or Replace Dataset against an ID that already exists, stops for the metrics owner with the evidence attached: "Dataset sales.pipeline.daily exists with 214 rows, last written 6 hours ago. Replace its schema? Widgets consuming it will need re-pointing." The person who owns the dashboards answers, because only a person knows who is watching them, and the approval is logged alongside the refresh history the pipeline already keeps.
Agent Capabilities
5 actionsAccount
1- Verify Authentication Pings the API to confirm the configured key is valid. The preflight check before any pipeline writes to a live dashboard.
Datasets
2- Create or Replace Dataset Creates a dataset with a typed schema, or replaces the schema of an existing dataset with the same ID. Fields carry a key, type, and display name, with currency for Money and time unit for Duration; Unique By marks the keys that make appends idempotent. The schema-first step every pipeline starts with, and a gated call when the ID already exists.
- Delete Dataset Permanently deletes a dataset and all its data; widgets built on it stop updating. Cannot be undone, which is why FlowRunner routes it through the metrics owner.
Data
2- Replace Data Replaces all rows in a dataset in a single call, values typed to the schema (numbers, YYYY-MM-DD dates, ISO 8601 datetimes, 0 to 1 percentages). The full-refresh verb for nightly syncs.
- Append Data Appends up to 500 rows per call without discarding existing data. Rows matching a Unique By key update in place, and Delete By rolls the oldest rows off at the retention limit. The verb behind self-maintaining trend datasets.
Frequently Asked Questions
What can FlowRunner do with Geckoboard?
FlowRunner agents can run Verify Authentication, Create or Replace Dataset, and Replace Data in Geckoboard, plus 2 more actions.
Does connecting Geckoboard to FlowRunner require OAuth?
No. Geckoboard connects to FlowRunner with basic authentication (a username and password), no OAuth flow required.
Can Geckoboard trigger a FlowRunner workflow automatically?
Geckoboard doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with Geckoboard
$100 in credits. No card required. Connect in minutes.