Back to directory

hypha

A TypeScript framework that pairs an agent-reasoning core with a production harness of event-backed execution, checkpoints, recovery and audit, so an application's specific behavior is declared separately from the runtime.

hypha

hypha is built around two cooperating layers: an Agent Core that handles reasoning, planning, tool selection and memory access, and a Production Harness that turns those decisions into bounded, event-backed execution with finite-state-machine control, policy and approval steps, checkpoints, recovery, replay and audit.

Product-specific behavior is declared through a DomainPack rather than hard-coded into the runtime: a DomainPack compiles an application's tasks, workflows, capabilities, prompts, memory, policy, evaluation and output contracts into the shared Core and Harness. A cross-cutting Cache and Reuse Plane then reuses validated reasoning, tool and memory results across runs, while remaining a disposable projection that can never replace event, artifact, receipt or checkpoint evidence as the source of truth. It is published as a set of aligned npm packages.

More in Agent frameworks

See all 902
  • Recurse

    recurse.run

    A serverless harness that lets a coding agent design, test, and deploy its own specialist sub-agents as tools, MCP servers, or bots, defined by a schema-validated manifest.

  • ProofForge

    shulgin.is-a.dev

    An AI-agent pipeline that decomposes a mathematics problem, proves each piece, and formalizes the result in Lean 4 and Mathlib until the Lean kernel accepts it, with six of its proofs already merged into Google DeepMind's formal-conjectures repository.

  • local-coder

    github.com

    Builds a local coding-agent team on top of Ollama and OpenCode by detecting the machine's hardware, recommending models for seven specialist roles, and testing whether each assigned model can actually use tools, edit a repository and pass verification.