Antfly, a Portland, Oregon-based startup, has secured $2 million in pre-seed funding to advance its retrieval engine for artificial intelligence agents. The investment round was led by Heavybit, with participation from 8-Bit Capital. Antfly also announced that its retrieval engine is now available for self-hosting, while organizations can join a waitlist for its fully managed cloud service.
Closing the Retrieval Gap
Modern AI agents depend on finding the right context at the right moment, but enterprise information remains scattered across structured and unstructured systems. Antfly describes this challenge as a retrieval gap that prevents many organizations from achieving the full value of AI outside coding tasks. The company aims to close that gap by unifying lexical, semantic, and graph retrieval in a single framework.
Replacing Fragmented Infrastructure
Organizations have typically assembled retrieval capabilities from vector databases, graph databases, rerankers, inference components, memory layers, and synchronization pipelines. This approach creates ongoing maintenance burdens and forces development teams to build the same glue code repeatedly across projects. Antfly replaces that fragmented stack with one engine that finds, ranks, secures, and refreshes context for AI agents.
Leadership and Founding Experience
Antfly was founded by James McDermott and AJ Roetker, who previously built data systems at Lytics at billion-profile scale. Their background in large-scale data infrastructure shapes the design and ergonomics of the retrieval engine. The founders argue that every AI team eventually builds the same retrieval stack, and they are offering a dedicated solution to eliminate that redundant work.
Investor Confidence
Heavybit General Partner Joe Ruscio noted that while AI has advanced coding significantly, broader enterprise adoption still lags because information is not indexed properly. He said Antfly addresses this issue with fine-grained access controls and deployment in the customer's own environment. Heavybit expressed confidence in the founding team's experience and its ability to lead from the earliest stage.
Early Adopter Experience
Visiting Media, an AI-powered hospitality platform for immersive selling, reported that building its own retrieval stack felt like pushing through mud. The company's CTO Eric Sniff said Antfly allowed the team to experiment with new indexes and different models within days. He added that no one maintains a retrieval pipeline anymore, and no one misses it.
Data Control and Security
Antfly is designed to leave organizations in control of their data and deployment choices. The platform includes fine-grained access controls so that sensitive information is only available to authorized agents and users. This approach addresses enterprise requirements for security and governance while still providing the performance benefits of a unified retrieval engine.
Availability and Next Steps
Organizations can self-host Antfly immediately by downloading the engine from the company's website. Those interested in the fully managed Antfly Cloud can join a waitlist for access through the company's website as the service becomes available. The company positions its retrieval engine as foundational infrastructure for future AI agents, similar to how databases became the foundation for applications.
With its new pre-seed funding and dual deployment options, Antfly is entering the market at a time when enterprises need reliable retrieval infrastructure for increasingly capable AI agents. The company's focus on security, governance, and self-hosting may appeal to organizations that require control over their data. As AI continues to expand beyond coding, Antfly's platform could become an important layer for connecting agents to the right information.