Master Poolside with our step-by-step tutorial, detailed feature walkthrough, and expert tips.
Explore the key features that make Poolside powerful for ai coding & development workflows.
Custom foundation models trained specifically for software engineering rather than general-purpose LLMs retrofitted for code. Models are deployed inside the client's environment and measured against evaluations co-designed with the client team, ensuring alignment with proprietary codebases and internal quality standards.
Deploy sophisticated coding AI on-premises for sensitive codebases — enabling regulated industries and defense organizations to use frontier models without data leaving organizational control
Single and multi-agent systems that can plan, use tools, and execute in sandboxed environments, all governed by explicit policies and end-to-end traces. Agents can collaborate on long-horizon tasks, with full auditability of tool calls, decisions, and outputs for compliance review.
Automate complex software engineering workflows like refactoring large legacy codebases, generating and running test suites, and coordinating multi-step feature delivery across repositories
Dedicated Poolside researchers embed with client teams to design, build, and operate intelligence inside the client environment. Unlike professional services, FDREs take joint outcome ownership — responsible for adoption, measurable business impact, and long-term operation, not just model delivery.
Get frontier-lab research talent working directly inside your engineering organization to build bespoke evaluations, agents, and workflows rather than adapting a generic product
On-premises, private VPC, and workstation (defense-only) deployment with role-based access control for both humans and agents. The platform runs inside multi-cloud, legacy, and air-gapped networks without rip-and-replace, and is co-designed to meet CISO and enterprise review board requirements.
Stand up enterprise AI coding infrastructure in regulated, classified, or air-gapped environments where SaaS coding tools are not permitted by security policy
IDE extensions, terminal binaries, and workflow integrations that travel with engineers, combined with connectors to repositories, databases, data warehouses, and private corpora within strict security boundaries. Designed to plug into existing toolchains rather than replace them.
Bring AI assistance into heterogeneous enterprise stacks — from monorepos to mainframes — while keeping proprietary data governed and traceable across the development lifecycle
Poolside operates at a fundamentally different layer than developer-facing assistants. Where Copilot and Cursor are SaaS IDE plugins running on shared cloud models, Poolside ships custom foundation models deployed inside the client's own infrastructure along with multi-agent orchestration, governance, and embedded research engineers. Devin is closer conceptually as an autonomous agent, but Devin is a hosted product with public pricing while Poolside is an enterprise platform with bespoke deployments and Forward Deployed Research Engineers who take joint responsibility for business outcomes. Based on our analysis of 870+ AI tools, Poolside is one of very few vendors combining foundation-model training, agent orchestration, and deployment services as a single enterprise offering.
Poolside does not publish pricing and operates exclusively on custom enterprise contracts. Given the $500M Series B funding, the Forward Deployed Research Engineer staffing model, and the on-premises foundation-model deployment footprint, engagements are typically six- to seven-figure multi-year commitments. There is no free tier, no trial, and no self-serve signup. Organizations interested in the platform engage through direct sales via the 'Talk to us today' contact on poolside.ai. Compared to per-seat tools like Copilot at $19–39/user/month, Poolside is structurally oriented toward large enterprises rather than individual developers.
Poolside supports three deployment modes: on-premises (in the client's own data center), private VPC (inside the client's cloud account on AWS, Azure, or GCP), and workstation-based installation, though the workstation option is restricted to defense clients. All modes keep data inside the client security boundary — Poolside explicitly states 'your data never leaves your control.' The platform is designed for heterogeneous environments including multi-cloud, legacy systems, and air-gapped networks, and includes role-based access control for both humans and agents, end-to-end traces, and governance aligned to CISO and enterprise review board requirements.
Forward Deployed Research Engineers (FDREs) are Poolside researchers who embed directly with client engineering teams to design, build, and operate AI systems tailored to the client's environment, codebases, and workflows. Unlike traditional vendor support or professional services, FDREs take joint responsibility for outcomes, adoption, and measurable business impact rather than just handing off a model. This hands-on delivery model is a core reason Poolside commands enterprise pricing — clients get dedicated research talent rather than a shrink-wrapped product. It also explains why engagements span months rather than hours and why the company targets high-consequence software environments.
Poolside was founded in 2023 by Jason Warner, former CTO of GitHub, and Eiso Kant, former founder of source{d}. The company's stated mission is 'for artificial general intelligence to drive abundance for humanity,' with software engineering chosen as the strategic beachhead on the path to AGI. The reasoning: software development requires understanding the world, multi-step complex reasoning, and long-horizon planning — capabilities that mirror human cognition and are well-suited to reinforcement learning. Poolside is headquartered in Paris with offices in the US and has raised roughly $626M in total, including a $500M Series B led by Bain Capital Ventures in October 2024 at a $3B valuation.
Now that you know how to use Poolside, it's time to put this knowledge into practice.
Sign up and follow the tutorial steps
Check pros, cons, and user feedback
See how it stacks against alternatives
Follow our tutorial and master this powerful ai coding & development tool in minutes.
Tutorial updated March 2026