Coactive AI Unveils Multimodal AI Platform Autumn '25, Transforming Content Discovery and Operations for Videos and Images

Audio By Carbonatix

SAN JOSE, Calif.--(BUSINESS WIRE)--Sep 30, 2025--

Coactive AI announced the availability of the Multimodal AI Platform Autumn '25 release (MAP Autumn '25), a major update that empowers companies to extract greater value from video and image content. Designed for enterprises managing large content libraries, MAP Autumn '25 streamlines content discovery, automates and enriches metadata tagging, and improves the ability to screen out negative results. Its flexible, API-first architecture helps developers to quickly build effective, content-based applications at low overall cost.

This press release features multimedia. View the full release here: https://www.businesswire.com/news/home/20250930205141/en/

Validate tags in minutes, not weeks. Auto Evaluator delivers precision, recall, and F1 quickly, so teams can deploy with confidence.

Unleashing the value of video and image content

Media organizations are racing to extract value from content libraries. With MAP Autumn '25, they gain precision content discovery and advanced metadata enrichment across video, images, and audio—fueling efficiency, revealing hidden insights, and protecting brand reputation.

“With MAP Autumn '25, we’ve reimagined how companies unlock value from their visual content,” said Cody Coleman, CEO, Coactive AI. “The Coactive MAP platform can power a wide range of content-driven applications, accelerate ad campaign production, and help create new customer experiences, all faster and more effectively than relying on foundation models alone.”

MAP Autumn '25 provides speed, accuracy, and ease of use for content intelligence

The re-architected Coactive platform enables a single embedding layer that powers multiple content use cases. This enables:

1. Faster, more accurate content discovery: Tiered video search results surface the most relevant matches first while using hybrid search signals—visuals and metadata—to deliver deeper recommendations and accurate identification of people, logos, scenes, moods, and dialogue.
2. Automated metadata tagging with higher accuracy: Achieves 81% zero-shot tagging accuracy, nearly doubling earlier benchmarks. Streamlined workflows, LLM-powered labeling, and ground truth reports accelerate tagging at scale.
3. Streamlined workflows, one-to-many leverage, and tagging at the shot, scene, and video level: Offers tagging at every level (shot, scene, video), intuitive result previews, and shareable tag templates for reuse across teams. This unlocks cross-department efficiency, reduces rework, and ensures a consistent metadata layer for downstream applications.

Key new features

• VIP identification: Easily search for or tag the very important people, products, or premises (concepts) relevant to your organization, with advanced named-entity recognition.
• Content safety controls: Built-in protections against unsafe, biased, or toxic content, ensuring brand suitability at scale.
• One-to-many workflows: Tag a small group of assets and apply results across multiple datasets, eliminating rework.
• Auto-evaluator: Assess visual tag performance at keyframe and shot levels, using high-quality ground truth generated on demand.

Powered by a fully rearchitected platform

With Coactive MAP Autumn '25, developers can select from best-of-breed foundation models to achieve their application goals. Optimize model selection for accuracy and performance, as well as freely select model weights. Coactive leverages a portfolio of leading models from OpenAI, Deepgram, Google Gemini, Alibaba Cloud (Qwen) and more.

In addition, Coactive enables multimodal fusion across multimodal signals. Use a variety of models to execute a complex set of tasks (chain of models), to benefit from a point model’s higher performance against sub-tasks.

Coactive MAP: a cost-effective way to maximize video content

MAP Autumn '25 sets a new standard for cost efficiency in video content processing. Designed to provide low total-cost-of-ownership, the platform enables customers to ingest up to 2,000 video hours per hour while reducing costs by 30–60% compared to traditional pipelines. Intelligent sampling cuts compute expenses by 6–10x, while cached outputs and tensors eliminate redundant processing, allowing organizations to leverage a single layer of embeddings without the need for costly reprocessing.

To find out more about MAP Autumn '25, please check out:

• “The Future of Video and Image Intelligence Is Here” by Coactive AI CEO, Cody Coleman
• “The Power to Create Video-Based Apps, Fast: How the Coactive Multimodal AI Platform Turns Months Into Minutes.” by Satya Komala, Head of Product and Engineering, Coactive AI

Learn more during this webinar: Unleash your video and image content: Accelerate content workflows with Coactive MAP ‘25.

About Coactive AI

Coactive AI transforms how organizations manage and monetize video and image content. Its Multimodal AI Platform (MAP) enables faster content discovery; automatically creates rich, broad metadata; and protects brand safety, while accelerating content workflows across the organization. Built for mid to large size enterprises, Coactive’s API-first, model-agnostic architecture integrates seamlessly into existing infrastructure to power a broad array of content-first use cases.

For more information, visit www.coactive.ai.

View source version on businesswire.com:https://www.businesswire.com/news/home/20250930205141/en/

CONTACT: Media Contact:

Jenna Valadez

Just Drive Media

[email protected]

KEYWORD: CALIFORNIA UNITED STATES NORTH AMERICA

INDUSTRY KEYWORD: DATA MANAGEMENT AUDIO/VIDEO APPS/APPLICATIONS TECHNOLOGY ARTIFICIAL INTELLIGENCE

SOURCE: Coactive AI

Copyright Business Wire 2025.

PUB: 09/30/2025 09:00 AM/DISC: 09/30/2025 08:59 AM

http://www.businesswire.com/news/home/20250930205141/en