• An advanced databasing system for the creative industries.

• We take different formats of creative data and incorporate them into unified databases.

• We build efficiency tools on top of that creative data powered by the most advanced new technologies available.

• Koobrik is YOUR institutional memory living on YOUR servers.

• Koobrik is a tool to triage your stack so you can make more effective choices about what to read next.

• Koobrik is a tool for finding scripts or projects that have been passed over or forgotten, giving new life to old dreams.

• Koobrik is not a replacement for reading a script.

• Koobrik is not a tool for writing a script.

• Koobrik does not tell you what projects to make and what projects to pass on.

• Koobrik is respectful of your intellectual property and does not use your IP in any way that you do not explicitly approve.

• Koobrik does not use any client’s data for the improvement of Koobrik’s services or it’s other client’s services.

• KOOBRIK DOES NOT TRAIN ON CLIENT IP UNLESS REQUESTED BY OUR CLIENTS FOR THEIR SERVICES ONLY.

Koobrik works with companies to organize their data so it is most useful to them. Taking advantage of new technologies in an ethical and respectful way.

Koobrik offers a suite of benefits designed to revolutionize how creative content, from screenplays and books, to contracts and beyond, is managed and discovered. By focusing on building efficiency tools atop a robust data schema, Koobrik ensures that creative works and their creators are given the attention they deserve, streamlining production processes and safeguarding against common industry pitfalls.

Koobrik saves all of your data entry time, freeing your team to get on with the process of making people’s creative dreams a reality.

Koobrik uses a combination of advanced tools to break down, database and categorize the elements of a piece of creativity.

• Koobrik uses classifiers like Google’s BERT model, AWS’s Comprehend, and Python’s NLTK library to identify characters, props, locations, vehicles etc.

• Koobrik uses embeddings models to allow client data to be searched semantically.

• Koobrik uses large language models (LLMs) to assist in plot summarizations, character breakdowns, etc.
  • Koobrik does not use client IP or unauthorized copyrighted material to train its own models.
  • We are very cognizant that many LLMs have been trained on pirated intellectual property (e.g., the books3 corpus). At its inception, Koobrik brought in Dr. Jonathan Frankle as its AI advisor. Dr. Frankle, is a well-respected thought leader in the world of AI and the Chief Scientist (Neural Networks) at Databricks. The key reason we asked him to be an advisor is because he is on the forefront of ensuring LLMs are ethically trained and respect IP rights. For some time, he has been working on a new model that has endeavored to avoid any copyright material. This will be the first tool of its kind and a major step in the improvement of these models.
  • Because of our unique relationship with Dr Frankle, we are given advanced access to their new model which has allowed us to prepare and test our systems to be ready for it.
• Koobrik uses multi-modal models like Google Gemini to classify visual data including video archive material, comic book material, and articles.

Koobrik works with its clients to make sure their data is safe and secure. If clients request that we host their data, we work with cloud providers like AWS, Google, Databricks, and Microsoft who offer several capabilities to support security and privacy requirements. These instances are in scope for common compliance standards such as Service and Organization Control (SOC), International Organization for Standardization (ISO), Health Insurance Portability and Accountability Act (HIPAA) eligible, and we can use them in compliance with General Data Protection Regulation (GPDR), and FedRAMP High. They are in the scope of the SOC 1, 2, 3 reports, allowing us to gain insights into their security protocols. They demonstrate compliance through extensive third-party audits of their controls. They are CSA Security and Trust Assurance and Risk (STAR) Level 2 certified, which validates the use of best practices and the security posture of their cloud offerings. With these services, our data and our client’s data cannot be used to improve the base models and is not shared with any model providers, which we think is CRUCIAL. We use Virtual Private links so we don’t have to expose our client’s data to internet traffic.

Where possible, we also work with various cloud providers to keep our clients’ data behind their own firewalls and security. In these scenarios, we would not have access to that data except when we are on location.

We like to keep our clients’ data where they want it, not where suits us. So, we are always open to discussing how to improve these systems in ways that make our clients most comfortable. We conduct security audits and only 4 people at our company have access to client material that is hosted on Koobrik’s servers.

Our advisors, investors, and partners DO NOT have access to client data.

While Koobrik is designed to support and enhance the creative industries through data management and efficiency tools, like any powerful technology, it has the potential to be used in ways that are contrary to its intended purpose. Recognizing these possibilities is crucial in implementing safeguards and policies to prevent misuse. Here are potential negative uses and the measures Koobrik takes to prevent them:

Potential Negative Uses:

If you want to remove your data or end your relationship with Koobrik, we will send you back your scripts and coverage and delete them from our cloud servers. It’s that simple.

If you decide to delete your account or remove any material from our platform, we adhere to the following protocol to ensure the secure and complete erasure of your data: