Memv – Memory for AI Agents

Name: Memv – Memory for AI Agents
Availability: InStock
Author: brgsk

by brgsk·Mar 30, 2026·4 points·3 comments

Visit Project View on HN

AI Analysis

●●SolidBig BrainNiche Gem

Predict-calibrate extraction reduces noise, but Zep and Mem0 already dominate the agent memory space.

Strengths

•Bi-temporal validity tracks fact truth versus learning time, solving agent drift issues.
•Predict-calibrate extraction minimizes storage noise by ignoring what the model predicted.
•Single db_url parameter swaps SQLite for Postgres without changing application logic.

Weaknesses

•Agent memory category is saturated with established players like Zep, Mem0, LangChain.
•Relies on external papers without benchmarks proving superiority over existing solutions yet.

Post Description

memv is an open-source Python library that gives AI agents persistent memory. Feed it conversations; it extracts knowledge.

The extraction mechanism is predict-calibrate (Nemori paper): given existing knowledge, it predicts what a new conversation should contain, then extracts only what the prediction missed.

v0.1.2 adds the production path: - PostgreSQL backend (pgvector for vectors, tsvector for text search, asyncpg pooling). Single db_url parameter — file path for SQLite, connection string for Postgres. - Embedding adapters: OpenAI, Voyage, Cohere, fastembed (local ONNX).

Other things it does: - Bi-temporal validity: event time (when was the fact true) + transaction time (when did we learn it), following Graphiti's model. - Hybrid retrieval: vector similarity + BM25 merged with Reciprocal Rank Fusion. - Episode segmentation: groups messages before extraction. - Contradiction handling: new facts invalidate old ones, with full audit trail.

Procedural memory (agents learning from past runs) is next, deferred until there's usage data.