How to Build an Automated Image Combiner Pipeline in Clay

According to data from lemlist's cold outreach studies, highly personalized email campaigns achieve a 17% reply rate, compared to a 7% reply rate for generic campaigns [Lemlist 2026 Survey]. This performance gap is driving commercial interest in scalable dynamic image generation sales workflows, as evidenced by a monthly search volume of 6,600 and a high cost-per-click of $4.06 for 'image combiner' keywords [DataForSEO Keyword Analysis]. An automated image combiner pipeline overlays variable dynamic graphics onto static background templates to produce customized outreach assets. While most guides detail manual tools like Canva instead of outlining scalable API layers integrated with GTM tables, this guide addresses the technical bridge needed to connect automated lead databases with cloud rendering tools.

Traditional design workflows rely heavily on manual human intervention. Graphic designers must open raster files, apply cropping parameters, and manually export vector formats. When managing brand assets for thousands of prospect companies, this manual approach creates a severe operational bottleneck. Growth operations teams frequently try to resolve this by saving raw assets to local folders, standard cloud storage accounts like Google Drive, Box, OneDrive, or Dropbox. These repositories lack version control, metadata schemas, and automated webhooks, making them poor destinations for dynamic pipelines.

Automating these steps eliminates manual graphic editing. By converting static prospect lists into dynamic assets, sales operations teams can automatically merge company logos, prospect names, and personalized messages into custom graphics. Establishing this pipeline requires connecting Clay, an outreach database and enrichment table, with dedicated image combiner APIs.

An automated image combiner pipeline relies on three main components: a structured database, an orchestration engine, and a rendering API. The pipeline processes raw data points, maps them to layout coordinates, and outputs finished graphics.

The first component is the static background template. This image serves as the canvas for the final asset, such as a laptop mockup, a customized billboard, or a personalized report cover. The graphic template remains identical across all outreach runs, providing consistent dimensions and visual context.

The second component is the coordinate and coordinate bounding system. To place dynamic layers like company logos or website screenshots on the background, the pipeline must pass exact coordinate parameters to the rendering API. The payload specifies the horizontal axis coordinate, vertical axis coordinate, scale bounds, and transparency settings for each dynamic layer.

The third component is the orchestration engine, which in this case is Clay. Clay acts as the data table that holds prospect information, including domain names, corporate logo URLs, and recipient names. For each row in the Clay table, the orchestration engine makes an HTTP call to the rendering API, supplying the specific coordinates, background template ID, and dynamic image URLs.

Managing the asset library, including static background designs and output graphics, requires a secure collaboration space. While raw file storage services like Amazon S3 or simple local drives are alternative options, they do not support team collaboration. Fast.io workspaces solve this coordination issue. Teams can store their design libraries and output assets in shared, org-owned folders. Granular permissions at the workspace and folder level ensure that automated agents can read and write assets without exposing sensitive files, while team members monitor the files in real time.

Clay University details how to execute custom HTTP request payloads and use formula fields to orchestrate dynamic creative generators [Clay University Documentation]. Since Clay does not have a native one-click integration for Bannerbear, you must configure Clay's HTTP API enrichment to send data to the Bannerbear rendering endpoint.

The first step is retrieving your Bannerbear credentials and setting up the background template. In your Bannerbear dashboard, create a new project and retrieve the Project API Key. Design your template, noting the exact name of the layers you want to make dynamic, such as 'company_logo' and 'prospect_text'.

Second, add the HTTP API enrichment to your Clay table. In the enrichment settings, select the manual configuration tab. Set the HTTP request method to POST and enter the Bannerbear image endpoint URL: https://api.bannerbear.com/v2/images. In the headers section, create an Authorization header and set its value to 'Bearer ' followed by your Project API Key.

Third, construct the JSON request body. The body maps the columns in your Clay table to the dynamic layers in the Bannerbear template. The JSON payload must follow this structure:

{
  "template": "your_bannerbear_template_uid",
  "modifications": [
    {
      "name": "company_logo",
      "image_url": "{{Company Logo URL}}"
    },
    {
      "name": "prospect_text",
      "text": "{{First Name}}"
    }
  ]
}

Because Bannerbear image generation is asynchronous by default, the API returns a status response containing a pending URL. To handle this, you can configure Clay to check the status after a short delay, or set up a webhook that writes the completed image URL back to your Clay table. When the image is generated, the API outputs a permanent URL for the merged asset.

Once the combiner pipeline outputs the generated graphics, the next challenge is storing, organizing, and auditing these assets. Generating thousands of custom images for campaigns can quickly clutter standard databases. Storing raw image links in spreadsheet cells makes visual inspection and quality control difficult.

While teams often write custom scripts to upload assets to local servers or Amazon S3 buckets, these systems are disconnected from the team's primary workspace. Fast.io provides an integrated environment where designers and automated agents manage these outputs. Developers can connect to Fast.io workspaces using a consolidated MCP toolset. Learn more about configuring these endpoints in our guide on storage for agents.

To process the generated files, you can use Fast.io's Metadata Views (/product/document-data-extraction/). Metadata Views turn unstructured document directories into a live, queryable database. Marketers can describe the fields they want extracted in plain language. The AI designs a typed schema, supporting Text, Integer, URL, and Date & Time formats. It scans the files in the workspace and populates a spreadsheet. In an image combiner pipeline, this extraction layer can read the generated PNG or JPEG files, extract the prospect's company name, evaluate the logo matching accuracy, and log the verification date. New columns can be added to the View without reprocessing the entire dataset.

Organizations on Fast.io operate on a paid subscription model with a 14-day free trial that requires a credit card to activate [Fast.io Pricing]. The platform offers three plans: Starter at $29/mo, Business at $99/mo, and Growth at $299/mo. Creating a user account is free, allowing agents to build workspaces and set up the pipeline before initiating an ownership transfer. The agent hands the configured workspace over to a human team member, who starts the trial on the pricing page to begin full outreach operations.

Consolidate and verify your dynamic outreach files

Store generated images, extract verification details with Metadata Views, and collaborate with your team in shared workspaces. Starts with a 14-day free trial.

Start 14-Day Trial

Scale operations require defensive design practices to handle missing data points and rendering failures. When running a pipeline across thousands of prospect records, variations in image files, empty fields, and API rate limits will occur.

One common issue is missing logo URLs. If a prospect's company logo is not found, sending a blank variable to the rendering API will cause the generation call to fail. You can prevent this by writing a JavaScript formula field in Clay. The formula checks if the logo URL column is empty and replaces it with a default fallback asset URL, ensuring the API always receives valid image parameters.

Another common problem is layout misalignment. A prospect's logo might have unusual dimensions, causing it to render outside the template coordinates. To resolve this, configure scale bounds and cropping properties within Bannerbear's layer settings. Setting scale bounds ensures that logos are scaled down to fit the coordinate box, preventing text overlap or clipping.

Large image payloads can also damage email deliverability. Dynamic images must be lightweight to prevent spam filters from flagging the outreach emails. To optimize performance, run a compression step in your middleware before writing files to your storage workspace.

If an automated script runs incorrectly and overwrites your layout templates, Fast.io's per-file version history provides a safety net. Designers can review the version list, restore previous background templates, and track changes across all assets. This history keeps concurrent agent edits auditable and secure.