Visual Intelligence Summit: Oct 22 in San Francisco Get your ticket (opens in a new tab)

Roboflow + Microsoft SQL Server

Insert detections, counts and pass or fail results from Roboflow Workflows as rows in a Microsoft SQL Server table, ready for reports, dashboards and other systems.

Microsoft SQL Server is Microsoft's relational database, used to store the records that reporting tools, dashboards and business systems query.

The Microsoft SQL Server Sink block inserts Workflow results into an existing table. You pass it a dictionary, or a list of dictionaries, whose keys match the table's column names, such as a timestamp, camera ID, class and confidence, and each one becomes a row. It connects on port 1433 by default and logs in with SQL Server Authentication, or with Windows Authentication when no user name is given.

How to connect Microsoft SQL Server to Roboflow

6 steps
  1. Set up an edge device

    Run your Workflows on a device managed by Deployment Manager, or on a self-hosted Inference server with enterprise blocks, on a network that can reach the SQL Server instance (port 1433 by default).

  2. Create the table

    Create a table with a column for each value the Workflow will send, such as timestamp, camera_id, class and confidence. The block inserts rows into an existing table and does not create one.

  3. Create a login for Roboflow

    Add a SQL Server login and database user with INSERT permission on the table, so the block can write rows without broader access.

  4. Shape the data

    In the Workflow, build a dictionary whose keys match the table's column names, for example with a Property Definition block, or a list of dictionaries to write several rows at once.

  5. Add the Microsoft SQL Server Sink

    Add the block and enter the host, port, database, table name, user name and password, then connect the data from the previous step.

  6. Check the rows

    Run the Workflow on a test image or stream and query the table to confirm new rows arrive with the values you expect.

What teams build with Microsoft SQL Server

  • Keep an inspection record for every part

    Insert a row for each inspection with the time, camera, result and confidence, so every decision can be traced later.

  • Report defects by shift

    Query the table from your reporting or dashboard tools to chart defects and output by line, shift or camera.

  • Feed systems that already read from SQL Server

    Put vision results in the same database your existing applications and reports already use.

Questions

No. The table must already exist, the column names must match the keys in the data, and the login needs INSERT permission on it.

Yes. Pass a list of dictionaries instead of a single one, and each dictionary is inserted as its own row.

SQL Server Authentication with a user name and password. If you leave both empty, the block uses Windows Authentication with the account the server runs under, which then needs permission on the database.

No. The Microsoft SQL Server Sink runs on your own device or self-hosted Inference server, which needs a network route to the database.

The Microsoft SQL Server Sink is part of the Enterprise plan. Talk to our team to add it to your workspace.

Why are companies choosing Roboflow?

The end-to-end platform for vision AI: create datasets, train models and deploy to production.

  • Flexible deployment modes

    Move from a scalable cloud API to edge hardware like NVIDIA Jetson without managing complex dependencies.

  • Integrated workflow builder

    Combine multiple AI models, custom business logic and third-party integrations into one production-ready pipeline in a low-code editor.

  • Model monitoring and alerts

    Dashboards track confidence, inference time and request volume for your deployed models, and email alerts tell your team when confidence suddenly drops.

View all
Get started

Build on the Platform

For developers, engineers, and technical founders who want to get hands on. Try the free tier; the docs are open.

Bring it into your operation

For heads of AI, operations leaders, and enterprise teams. Bring a known problem, or work with us to find the first one worth taking on.