Glean launches upgraded Glean Agents toolkit with new development tools – News

Glean Technologies Inc. today moved its Glean Agents toolkit into general availability with new features designed to ease developers’ work.

San Francisco-based Glean is backed by more than $600 million in funding. It provides an artificial intelligence platform that helps workers search for information across their company’s spreadsheets, the public web and other sources. It also offers other AI features that automate tasks such as summarizing documents.

Glean Agents, the offering that become generally available today, made its initial debut in February. It provides tools that promise to help developers more quickly build AI agents. There are also prepackaged AI agents for common business tasks. It’s rolling out enhancements across both those feature sets.

The first addition to the toolkit, the Glean Agents API, provides software building blocks for creating custom AI agents. It also eases the task of enabling AI agents to exchange data with one another while collaborating on tasks. On launch, there are client libraries that enable developers to interact with the API using the TypeScript, Go, Java and Python programming languages.

Developers can optionally equip their AI agents with search features by connecting them to Glean Search and Glean Assistant. The former tool provides Google-like information retrieval features, while the latter also covers certain other use cases such as business data analysis. AI agents can be integrated with Glean Search and Glean Assistant through an MCP server that made its debut today.

Another new feature, the Glean model hub, allows developers to connect their AI agents to popular third-party large language models. On launch, there’s support for 15 LLMs from Amazon Bedrock, Google Vertex AI and Azure OpenAI. A so-called universal model key allows customers to gain access to all the LLMs at once instead of having to procure them separately.

“This eliminates the complexity and cost of individually obtaining keys and using models across different providers,” Glean co-founder and Chief Executive Officer Arvind Jain wrote in a blog post today. “It also allows you to mix and match models at every stage of your agentic workflow.”

In addition to speeding up the process building AI agents, Glean Agents now promises to help customers secure them. The toolkit is gaining a module called Glean Protect that allows companies to regulate who can develop and build AI agents, as well as how they may be shared. When AI agents interact with business data, the tool ensures that they don’t access sensitive records.

Enterprises with particular stringent cybersecurity requirements can now deploy Glean’s software on-premises. According to the company, this deployment option is made possible by a partnership with Dell Technologies Inc. Customers can run both Glean’s platform and the AI agent in their internal data centers.

“As actions gradually evolve from being manually triggered to subscription-based automation with agents running actions automatically in the background, Protect also pauses or cancels an agent at any time its scope changes,” Jain detailed.

Glean Agents is moving into general availability with more than 30 pre-built agents. Some automate department-specific tasks such as processing technical support tickets, while focus on personal productivity. Additionally, there’s a structured query agent that makes it possible to run analyses on data hosted in Snowflake data using natural language prompts.

Image: Glean