Agent orchestration
Poolside offers single-agent and multi-agent systems that can plan work, call tools, and execute in sandboxed environments with governed policies and full traces.
Poolside is a foundation model and agent platform for software work, with terminal, IDE, browser, and headless deployment for governed enterprise use.
Poolside is a foundation model company focused on software agents and enterprise systems for deploying them. The product suite includes models, agent workflows, and deployment surfaces designed to fit into the tools and environments teams already use.
The company positions Poolside around high-consequence software work and says its systems can run in sandboxes, connect to internal data sources, and operate within a customer's security boundary. Deployment options mentioned on the site include on-premises, a VPC, and workstations for defense-only use.
Poolside offers single-agent and multi-agent systems that can plan work, call tools, and execute in sandboxed environments with governed policies and full traces.
Agents are available in terminal, editor, browser, and headless workflows, so teams can use the same platform from shell commands, IDEs, CI/CD, or a web interface.
Poolside connects to repositories, databases, data warehouses, and private corpora while keeping access inside strict boundaries.
The models can be deployed in your environment and evaluated against benchmarks created jointly with your team, supporting enterprise-specific validation.
Fine-grained permissions and versioned checkpoints let teams control agent actions and review, roll back, or forward agent work alongside existing Git processes.
Forward Deployed Research Engineers work with customer teams to design, build, and operate intelligence where the work happens.
Use Poolside when engineers want an agent that can run from the shell, work across repositories, and feed results into scripts or automated workflows.
Use it inside an editor when teams want inline suggestions, edits, and context without switching between tools or windows.
Use the browser or headless modes for teams that want a full interface for ad hoc work or CI/CD automation for code review, generation, and testing.
Use the enterprise deployment model when software work must stay inside strict security boundaries, including on-premises, VPC, or restricted workstation environments.
Use Poolside when teams need agents to connect to internal repositories, databases, warehouses, or private corpora while keeping access governed and traceable.
Poolside provides foundation models and an agent platform for software work, with deployment options in your environment and surfaces such as terminal, IDE extensions, browser access, and headless workflows.
The source describes agents that can plan, use tools, and execute in sandboxed environments, with governed access and end-to-end traces.
Poolside says its models are trained in-house and that Laguna XS.2 is open-weight under Apache 2.0. Laguna M.1 and XS.2 are available through the dedicated API and on OpenRouter.
Poolside emphasizes deployment inside a customer's security boundary, including on-premises, in a VPC, or on workstations for defense-only use.