Back to directory

SlideRule

SlideRule is a product-rehearsal engine that takes one sentence describing an idea and runs it through a gated pipeline of spec, pages, structure, bind and closure steps to produce a runnable application preview.

SlideRule

SlideRule turns a one-sentence product idea into a gated rehearsal of a working application rather than a single-shot generation. Its factory runs a fixed sequence of hops, spec, pages, structure, bind and closure, and a hop that does not actually change a deliverable is not allowed to report success; a typed control plane of scope cards, assumption cards and hop intent keeps later steps from silently overriding earlier decisions, and missing evidence is always treated as missing rather than assumed fine.

The generated application runs as a browser app described by a five-system JSON schema covering pages, role-based access, workflow, data and generative content, with no backend needed for the preview; a Python FastAPI service is the authoritative engine behind it, with a thin Node proxy in front. Architecture compilers check the generated code itself, failing on undeclared dependency edges, new import cycles, or a growing ratio of code living directly in function bodies rather than proper modules. A static, zero-install demo lets a visitor try a captured end-to-end rehearsal and run a finished example app directly in the browser.

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.