We've standardized the way you control insight generation across our APIs by improving our insightSubtypes parameter on the generate-insights endpoint. This brings the same granular insight control you're familiar with from generate-content endpoint to our generate-insights endpoint.

Feature Details

You can now specify which insights you want using the insightSubtypes array parameter in generate-insights requests, just like you do in generate-content:

{
  "insightSubtypes": [
    "companyMarketTrends",
    "behavioralProfile",
    "socialMediaProspectRecentLinkedInPosts"
  ],
  "contactEmail": "daniel@acme.com"
}

The API will return only the requested insight types in the response:

{
  "insights": [
    {
      "type": "companyMarketTrends",
      "value": "Acme Corp is experiencing 40% YoY growth in the enterprise segment...",
      "confidence": 0.89
    },
    {
      "type": "behavioralProfile",
      "value": "Decision maker who prioritizes ROI and technical excellence...",
      "confidence": 0.85
    },
    {
      "type": "socialMediaProspectRecentLinkedInPosts",
      "value": "Recently shared thoughts on AI adoption in manufacturing...",
      "confidence": 0.95
    }
  ]
}

Availability

This standardization is available in:

• generate-insights v3.2 and newer

We've added support for Gemini Flash 1.5 as a new model option while deprecating fine-tuned models across all endpoints. This update brings ultra-fast generation capabilities to our API while streamlining our model offerings.

New Model Details

The Gemini Flash 1.5 model offers:

• 1-3 second response latency
• High-quality output comparable to base models
• Optimized for high-volume requests and real-time applications

Example usage in generate-content

{
  "model": "gemini_flash_1.5"
}

Important Deprecation Notice

All fine-tuned model endpoints have been deprecated. Requests to these endpoints will now return an error. Please migrate your applications to use our base models instead.

Availability

This update is available on all active endpoints.

We've standardized our insight taxonomy to make it more intuitive and powerful. You can now use the same subtype names across both our generate-insights and generate-content APIs. Each request can include multiple subtypes, giving you more precise control over which insights appear in your responses.

Check out our complete list of available subtypes to see what's possible. Here's a quick example of using multiple subtypes in a single request:

{
  "enabledInsights": [
    "companyMarketTrends",
    "behavioralProfile",
    "socialMediaProspectRecentLinkedInPosts"
  ]
}

This standardization makes our APIs more predictable and easier to use while giving you more granular control over your insight generation.

Availability

This functionality is available in the following endpoints and onwards:

• generate-content v3.6
• generate-insights v3.2

Insight Control is a feature that allows you to manage which types of insights are included in responses from the generate-insights API.

As of today, only the insights toggled on will show when making a generate-insights query, assuming insightSubtype isn't used as a parameter.

By enabling or disabling specific insights, you can tailor results to align with your strategy.

1. Default Behavior: Retrieve Top 20 Insights

By default, when no insightSubtype is specified, the generate-insights API will return up to 20 insights ranked by relevance, but only from the types you have enabled in Insight Control.

2. Retrieve Insights by Subtype

If an insightSubtype is explicitly specified in your request (e.g., "podcast" or "financial"), the API will attempt to retrieve and return those insights even if the subtype is disabled in your settings.

Autobound's generate-insights API now offers enhanced precision with a new insightSubtype parameter, allowing users to request specific insight types for their contacts.

This new feature streamlines data retrieval by focusing on the exact insights users need, such as "Hiring SDRs" or "Negative quarterly revenue growth." Without specifying a subtype, the API will continue to return a ranked list of the 20 most relevant insights for the contact.

This update provides greater control and flexibility, enabling users to tailor their data requests for more targeted and effective sales outreach. Developers can easily implement this new parameter using a simple string input, and comprehensive error handling ensures smooth integration.

For a complete list of valid subtypes, refer to this Google Sheet.

Welcome to the developer hub and documentation for Autobound API! You can find info on recent API and schema improvements here.