Perseus

ActiveAI-native

Applied Lab focused on Search in Latent Spaces.

What it does

Perseus takes an iterative branching approach to search in high dimensional spaces. Our first product helps coding agents find the right codebase context before they write. It searches files, symbols, and dependencies in parallel, so models can make better changes with less trial and error. Perseus Search gives Mythos-level programming ability to Opus-level coding agents.

Its site says now · captured 2026-08-23

Perseus · Codebase search at perseus speeds - Perseus

Perseus is the retrieval engine that grounds coding agents in real code: multi-hop search across your codebase, answers with exact file-and-line citations, warm in about a second. Built by Efficient Systems.

Next to it in Agent infrastructure · Search and retrieval for coding agents

  • AmikaY Combinator F25

    Infra to build your own software factory

  • MossY Combinator F25

    Real-time semantic search for Conversational AI

  • SpecificY Combinator F25

    AWS for coding agents

  • stagewiseY Combinator S25

    The Open Source Agentic IDE

  • ZenbuY Combinator X26

    The extensible IDE for coding agents

  • AirweaveY Combinator X25

    The open-source context retrieval layer for agents

  • MorphY Combinator S23

    Fast Models Optimized for Coding Agents

  • Scott AIY Combinator F25

    The planning layer for teams that build with coding agents