Azure Table Storage

Connect AI agents to Azure Table Storage, the NoSQL key/attribute store in Azure Storage. Agents manage tables and entities keyed by PartitionKey and RowKey with OData filter queries.

12 actions available

A stream of schemaless records arrives from an upstream service

Agent runs Insert-Or-Replace Entity to sync each record idempotently

Agent reads back entities with Query Entities using an OData filter

Agent identifies which entities still need handling

Agent runs Merge Entity to mark handled records without touching other fields

Agent alerts the team of the batch summary via Slack

Any Delete Table pauses for approval

What This Integration Enables

Agents persist high-volume schemaless records in a low-cost store, look up and filter entities by PartitionKey, RowKey, or custom properties with OData `$filter` and `$select`, write idempotently with insert-or-replace and insert-or-merge upserts, and manage table lifecycle as part of a data pipeline. Every entity is uniquely identified by a PartitionKey and RowKey pair; other properties are schemaless and typed automatically. Replace operations write the whole entity so omitted properties are removed, while merge operations touch only the supplied properties and retain the rest.

Without FlowRunner

Key never protected Ad hoc scripts pass the account key around to reach the store

Non-idempotent syncs Re-running a sync creates duplicate rows

Replace overwrites An update removes properties it did not mean to touch

With FlowRunner

Key stays in the runtime SharedKeyLite signing happens locally; the key never leaves

Idempotent by upsert Insert-Or-Replace and Insert-Or-Merge make syncs safe to repeat

Merge preserves fields Merge Entity touches only supplied properties and retains the rest

Use Case Scenarios

Idempotent record sync

Records arrive from an upstream source, for example new rows from a spreadsheet trigger. The agent calls Insert-Or-Replace Entity keyed by PartitionKey and RowKey, so each record syncs exactly once no matter how often the flow runs. A durable, low-cost record of every event builds up automatically.

Filter, act, and mark handled

The agent calls Query Entities with an OData `$filter` to pull unprocessed records, alerts the team on each with Slack, then calls Merge Entity to mark the record handled without disturbing its other properties. The queue drains cleanly, and no field is lost to an accidental overwrite.

Table teardown with a human gate

At the end of a pipeline's life, an agent is asked to remove a table. Before it runs Delete Table, it does not act on its own. It routes the table name and a summary for approval, and deletes it only after a person confirms.

Human-in-Loop Highlight

Individual entity writes in Table Storage are cheap and reversible; deleting a whole table is neither, because it removes every entity in one call. FlowRunner's answer is human-in-the-loop, an execution pattern where the agent pauses on its own, assembles the context and the choices, routes to a human on their preferred channel, and resumes the moment they respond. When a flow reaches Delete Table on a table flagged as production, the agent pauses before executing and asks through Slack: "Ready to Delete Table `telemetry`. This removes all of its entities. Approve or cancel?" The delete runs only after a person confirms, with the approver and timestamp captured in the run log. A connector can drop any table; an orchestration layer knows which drops should stop and ask.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via Slack

Human decides

Agent resumes with decision

Agent Capabilities

12 actions

Tables

Create Table Creates a table in the storage account.
Delete Table Permanently deletes a table and all of its entities.
List Tables Lists the tables in the storage account, paged via a continuation token.
Query Tables Queries tables by name with OData filter syntax.

Entities

Delete Entity Deletes one entity by PartitionKey and RowKey.
Get Entity Fetches one entity by PartitionKey and RowKey.
Insert Entity Inserts one entity keyed by PartitionKey and RowKey; fails if the entity already exists.
Insert-Or-Merge Entity Inserts an entity, or merges the supplied properties into an existing one, for idempotent writes that retain other properties.
Insert-Or-Replace Entity Inserts an entity, or replaces an existing one entirely, for idempotent writes where omitted properties are removed.
Merge Entity Updates only the supplied properties of an existing entity, retaining the rest.
Query Entities Filters entities by PartitionKey, RowKey, or custom properties using OData `$filter` and `$select`, paged via continuation tokens.
Update Entity (Replace) Replaces an existing entity entirely; properties you omit are removed.

Start building with Azure Table Storage

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us