Back to directory

disp8ch

A self-hosted AI workspace that turns a plain-English chat request into an editable visual workflow, with a planner, implementer and independent reviewer completing an acceptance contract before any side effect runs.

disp8ch

disp8ch turns a chat request, such as building a daily research digest and sending it automatically, into a real, editable workflow on a visual canvas of triggers, nodes and schedules, and asks for approval before saving it. It can also record a product discussion and turn its cited decisions into assigned work, or run a planner, an implementer and an independent reviewer through an acceptance contract before publishing a result, with the underlying evidence available for inspection.

Everything runs self-hosted and works with fully local models with no API key needed, and every side effect is described as approval-gated and auditable, with no hidden prompt chains or silent writes. Its foundation connects a chat surface, visual workflows, boards, a hierarchy view, a multi-agent council, data sources, memory, skills, a design studio, channels, and an approvals and activity log; a later addition adds consent-gated live sessions with timestamp-cited delivery briefs, bounded repair with executable proof, and outcome receipts that keep a decision, its owners, evidence, cost, approval and public result connected.

It can use large third-party models through an operator's authenticated coding-assistant session as the planner and independent reviewer, while a separate, faster or cheaper configured model performs implementation; the application does not copy or expose that session's authentication token.

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.