The AI agent platform for the modern enterprise.

Libra WorkBase is Libra’s enterprise workspace for connecting company knowledge, tools, people and workflows. It provides a permission-aware context layer across the systems teams rely on and supports governed execution of work across those systems.

Notion, Confluence, Jira, Linear, and Asana are native, alongside Slack, Google Drive, and more. Every connected source is unified into one searchable index that stays current as your team works.

Yes. Libra is designed to connect processes, handoffs, decisions and dependencies across teams. Recurring processes can be turned into workflows that track ownership and move work forward with governed automation.

Libra and Glean both address enterprise knowledge and AI, but Libra is positioned around both knowledge and execution. Libra connects company context to AI agents and workflows, with deployment options that include managed cloud, private VPC and self-hosted environments.

Libra is designed around connecting company context to the systems where work actually happens. Instead of focusing only on conversational assistance, Libra combines a permission-aware company context layer with agents that can plan and execute multi-step work across connected tools.

Yes. On Enterprise, the full services stack (server, worker, database, and cache) runs inside your own environment (Bring Your Own Machine), while Libra manages the harness, models, and worker engine. Indexed content never leaves your walls.

Libra offers three deployment approaches: managed cloud, private VPC and fully self-hosted deployment. This gives enterprises a choice between a managed experience and greater control over where the Libra services and indexed company content run.

Yes. Libra supports enterprise deployment models that give organizations control over where their Libra environment and company data are hosted. Depending on the deployment model, customers can use managed cloud, a private VPC or a fully self-hosted environment to meet data-residency requirements.

Libra is designed to be permission-aware across the systems it connects to. Its knowledge layer inherits access from source systems so information and actions can respect the permissions a user already has, rather than relying on a separate parallel permission model.

Yes. Libra supports role-based access control for enterprise access management. RBAC can be used alongside permission-aware access so administrators can control access while the underlying source-system permissions continue to govern what users can retrieve and act on.

Yes. Libra provides auditability for agent activity and governed actions, helping enterprises understand what work was performed, what systems were involved and where human approvals or interventions occurred.

Yes. Libra is SOC 2 Type II compliant, providing an independently assessed security framework for the controls used to protect customer information and support enterprise security requirements.

Yes. Libra encrypts data in transit and at rest as part of its enterprise security controls, helping protect company information while it is being transferred between systems and while it is stored.

No. Your content is never used to train foundation models. It's indexed solely to answer your team's questions and run your agents, under SOC 2 Type II controls.

Each seat includes a monthly pool of credits. One chat answer uses 1 credit; one agent run uses 3, since it reasons across sources and executes steps. Additional credits are $1 each, and unused credits don't roll over.

Yes: annual and multi-year contracts come with custom pricing. Reach out to sales and we'll put together terms that fit your rollout.

Every paid plan starts with a 30-day free trial, no card required. For a scoped Enterprise pilot with self-hosting or SSO, talk to sales and we'll stand up an evaluation environment.