← Back to work
Local Multi-Agent LLM Platform
A modular, offline-capable multi-agent platform — OpenAI-compatible gateway, control plane, autonomy engine, RAG, and IDE integration — with an automated adversarial peer-review step and zero hosted-LLM dependency.
- automated
- adversarial peer-review before merge measured
- ~32k
- lines of code estimated
- 0
- hosted-LLM dependencies measured runs fully local / offline
Private infrastructure — not published.
Context
I wanted a senior/junior agent team I could run entirely on my own hardware — no cloud, no per-token bill, no data leaving the building — that could actually maintain and extend itself under review.
What I built
- An OpenAI-compatible gateway fronting two locally-served models, with a control plane and a browser dashboard for full visibility.
- Automated adversarial peer-review — a generator reviews recent commits before merge, with a separate multi-dimension review recipe (correctness, security, performance, API-contract) for larger changes — so generated work is challenged before it is trusted.
- An autonomy engine with an allow / approve / block policy gating every agent action, least-privilege fail-closed tool bridges, and per-project privacy isolation.
- A deterministic competence benchmark (keyword-rubric, no LLM judge) to measure and gate quality rather than assert it.
Why it matters
It keeps working, catches its own mistakes before merge, and never depends on a hosted model to run.
The platform itself is private infrastructure and is not published; the RAG and retrieval components extracted from it are the public repositories linked elsewhere on this site.