Imbue

Imbue is an AI Research Labs research and product company that builds AI agents intended to work for humans, with public open-source tools available at $0 and custom enterprise or research pricing not publicly listed on the provided website content. It is best suited for researchers, developer-tool builders, and technical teams evaluating reasoning-focused AI agents rather than buyers looking for a conventional self-serve SaaS product today.

The provided Imbue website content states the company’s core positioning directly: “We build AI that works for humans.” Combined with the existing listing data, Imbue is not presented as a conventional chatbot vendor or workflow automation app; it is an AI research lab and product company focused on agents that help people think, create, and build. The current listing identifies Imbue as founded by Kanjun Qiu and Josh Albrecht, and describes its work as spanning openly released developer artifacts, proprietary model training, and reasoning-heavy agentic workloads. The public About page says Imbue was founded in 2021, identifies Kanjun Qiu as CEO and co-founder and Josh Albrecht as CTO and co-founder, and describes a mission to build tools that put software under user control.

Compared to the other AI Research Labs tools in our directory, Imbue appears more research-led and less commercially packaged than API-first platforms that publish model catalogs, usage-based pricing, and broad product documentation. Its strengths are clearest for teams studying AI agents, reasoning evaluations, datasets, and tooling, especially where open technical artifacts can be useful without depending on a general-availability hosted model API. Verifiable public facts include the June 25, 2024 research series on a 70B-parameter model; the public product list covering mngr, Bouncer, Sculptor, Blueprint, Vet, and Latchkey; the September 26, 2025 Sculptor product post; the April 30, 2026 Sculptor update; and 2026 posts including Bouncer on April 8, mngr on April 2, Vet on March 5, Latchkey on February 27, and Beating ARC-AGI-2 with Code Evolution on February 27.

Based on our analysis of 870+ AI tools, Imbue fits best as a frontier-style AI lab to watch or learn from and as a source of open agent tooling rather than as a plug-and-play business application. The visible website content does not provide customer counts, integration counts, public uptime commitments, enterprise seat prices, API conversion rates, paid plan names, usage limits, pilot fees, minimum contract values, or a full commercial pricing table, so those details should not be assumed. For professional directory users, the practical takeaway is that Imbue is compelling if your priority is human-centered AI agent research, open-source agent infrastructure, and technical insight, but it is currently harder to evaluate as a production vendor because commercial packaging remains limited and paid procurement requires a direct scoping conversation.