.png)
Clairo AI offers private, customised, and sustainable Generative AI solutions to businesses and organisations across various sectors. Clairo AI is designed to boost productivity, improve decision-making, and fuel innovation. Clairo’s state of the art platform allows you to empower your workforce using an AI-powered assistant, automating customer support, content creation, troubleshooting, all while processing large quantities of data to continuously improve application performance.A subscription to Clairo AI allows you to input proprietary data into our secure platform, create customised large language models for your business, and benefit from generative AI solutions without compromising on cost or data privacy.
Clairo AI’s platform is supported by all devices, accommodating a wide range of users. You can access the platform via standard web browser, tablet, laptops of all sizes, and mobile devices. Our platform is also compatible with most major operating systems, including Windows, macOS, and Linux.
Clairo AI supports both open source and closed source Large Language Models, making us model agnostic. Open source models are self-hosted on our Clairo AI platform, whereas the closed source models are accessed through an API. The following is a list of the models that we support.
Open Source:
- Llama3-8B and Llama3-70B. Provider: Meta AI
- Mixtral 8x7B and Mistral 8x22B. Provider: Mistral AI
Closed Source:
- GPT-3.5-Turbo and GPT-4-Turbo. Provider: OpenAI
- Mistral Large. Provider: Mistral
As we grow, we plan to support many more models, including the Anthropic Claude models, Cohere Models, Databricks, etc.
Clairo AI has established a strategic partnership with a high-performance compute (HPC) center in Iceland, Responsible Compute, to achieve our net-zero goals and develop our Green AI platform. This center harnesses the power of hydro and geothermal energy sources, which are renewable and sustainable, to fuel its infrastructure, significantly reducing the carbon footprint associated with High-Performance Computing.
Clairo AI's large language models (LLMs) are operated on Responsible Compute’s servers located in Iceland, ensuring that our models are powered entirely by renewable energy. This approach not only delivers high-performance computing but also minimises our impact on the environment, aligning with our commitment to sustainability and objective to achieve 100% net-zero operations.
Read more about Generative AI's Impact on the Environment and Clairo's Commitment to Green AI here: https://www.clairo.ai/blog/green-generative-ai
Clairo AI has two options for deployment of your AI models and agents. The first option is to host on the Clairo AI Platform, which is our preferred choice. The Clairo AI Platform is run on our partners servers, Responsible Compute, in Iceland. The data centre is highly secure and closed off from the internet. All your data and models are protected behind a firewall. An alternative second option is to self-host the Clairo Platform and models in your environment for an added layer of protection. For this, we support deployments in AWS, Azure, GCP and on-premises. Our AI Engineers will assist you in understanding the cost implications around different infrastructure. Clairo AI will then manage your agents and models in your own closed environment.
Closed-loop refers to closed loop AI, which is built into our name: C-l-ai-ro. Generative AI has significantly shifted the way businesses think about their data. Businesses can now leverage their large data assets to provide value for their employees and customers. Clairo AI was founded on the basis that companies want full protection of their data and AI models. At all times, businesses should have peace of mind that their data is not moving around the internet to different APIs. Clairo AI builds your Generative AI solutions using open-source models that are hosted in a closed environment.
Following our Beta program, Clairo AI will move to fixed pricing for customers. Based on the number of messages you have on average per quarter you will be placed in a pricing tier. This fixed pricing allows companies to accurately forecast the AI expenses, without worrying about variability in usage. If demand for your AI changes on average over a quarter, we’ll shift you into different pricing tier, with enough warning.
Bias mitigation has two levers. The first lever is the model and the second is the dataset. Clairo AI only deploys open-source models that have been checked for bias. For example, Meta AI’s Llama models are well supported and would have been deeply checked. In addition, Clairo AI’s data scientists will evaluate your datasets to determine potential areas of bias. Any piece of data that may cause bias will be removed from the dataset. Clairo AI also has a feature called “libraries”, which allows companies to group certain documents together. If a library of documents is causing bias, Clairo AI allows you to simply remove that library from the Agents knowledge base.
An AI Agent in the context of Generative AI is a specialised model designed to perform a specific task or job, equipped with necessary algorithms and knowledge. In a corporate setting, multiple AI agents can be employed to enhance productivity and streamline operations, serving various roles such as customer service, data analysis, content generation, or personalised recommendations. These agents, like AI chatbots for customer service or AI models for data analysis, leverage AI's power to automate tasks, improve efficiency, and augment human capabilities, showcasing AI's transformative potential in the modern business landscape.
Clairo AI defines a token as a text piece that can be processed by the model. The token limit you define in the platform is used as a parameter for defining the maximum number of text pieces that the model will process or generate in a single instance.
To contact a member of the support team, simply email us at support@clairo.ai or send us a message on LinkedIn at https://www.linkedin.com/company/clairo-ai/
