Conclave Personal: open-source multi-agent verification tool assigns roles to LLMs for self-correction
HospitalSlight7930 · reddit · 2026-08-13
Developer released Conclave Personal v0.1.5 Alpha, a multi-agent verification tool that brings multiple models into a shared conversation with explicit roles: Writer drafts, Critic challenges, Judge evaluates. Every execution is traceable as a 'run' with status, errors, duration, and token usage. Conversations, agents, and workspace remain local; Ollama runs without API key. The code was entirely written by language models, embodying the multi-agent verification concept.
Related event: Open-Source Multi-Agent Tool Conclave Enables LLM Cross-Validation(2 posts)→
More from coding & agent
- Dev Uses Codex to Revamp three.js Boids Demo with Van Gogh Style — simonxxoo · 2026-08-13
- Open-source Framework Harness Hits 7k Stars Rapidly — vista8 · 2026-08-13
- GRADE: Optimizing Multi-Agent Inference Costs via Gated Routing — Tanmoy_Chak · 2026-08-13
- Open Source Tool Grok Build GUI Supports VS Code and Multiple Models — PawelHuryn · 2026-08-13
- Roboflow Engineer to Share Real-Time Video CV with Supervision at Lisbon AI — dscape · 2026-08-13
- Caveman Optimizes Agent Context Representation, Slashing Input Tokens by 33% — VeryVexxy · 2026-08-13