Déjà Review: An Open-Source Code Reviewer That Remembers Your Past Decisions
Ill_Pressure5506 · reddit · 2026-09-29
A team open-sourced Déjà Review, a code-reviewing agent with persistent per-project memory built on Hindsight, with a GitHub repo and demo video.
- Memory: each project gets its own Hindsight memory bank; before reviewing, the agent recalls team conventions and prior review decisions
- Structured findings: the LLM generates structured review findings via Pydantic models
- Filter rejected suggestions: accept/reject outcomes are stored with retain, so previously rejected suggestions get filtered from future reviews, cutting repeat nitpicks
- Validation: review evidence and memory references are validated before reaching the UI
The stated goal: not remembering everything, but remembering decisions that matter.
Related event: Team Open-Sources Déjà Review, a Code Review Agent with Persistent Memory(2 posts)→
More from coding & agent
- Matt Shumer says Opus 5.5 cleared a 3D printer bed after 16 tries, taking humans out of the loop — mattshumer_ · 2026-09-29
- Claude autonomously builds a 'p(doom)' Beat Saber clone in one afternoon — burny_tech · 2026-09-29
- AlignOPSD fixes decision-timestamp mismatch in agent distillation, beating GRPO by 5.5-8.7% — Mingju Chen · 2026-09-29
- CMU's ControlScope: how much of a running agent workflow should be revised? — CarnegieMellonU · 2026-09-29
- CompoWorld scales general agent training by composing environments from reusable services — AllSpark-Research · 2026-09-29
- Adaptive Consistency Graph lifts long-horizon agent success from 44.5% to 50.2% — Beihang · 2026-09-29