compaction

ToplingDB is a cloud native LSM Key-Value Store with searchable compression algo and distributed compaction

This is an implementation of RFC 6902 JSON Patch written in Java

Go library for protobuf compatible sstables, a skiplist, a recordio format and other database building blocks like a write-ahead log. Ships now with an embedded key-value store.

FirstCoder is an open-source local coding agent that’s useful enough to run, small enough to read end to end, and structured enough to learn from. If you’re building your own agent, it’s a practical reference—not a black box.

"Work appears completed or recorded. Run /compact to save tokens."

Compaction runtime for Apache Iceberg.

Global Scalable Paleo Landscape Evolution Model

Claude Code skill for session handoffs — survives context compaction, chain-links across sessions.

CliffCompaction — autocompaction for long-horizon coding agents

Claude Code's compaction engine, as a drop-in DeepAgents middleware

Pacman: An Efficient Compaction Approach for Log-Structured Key-Value Store on Persistent Memory

Use OpenAI Responses API native compaction in OpenCode

Python wrapper of Sparrow, State-of-the-art Nesting for 2D irregular strip packing

Long-term memory for Claude Code. Indexes every session as SQLite+FTS5, exposes recall via MCP, and ships PreCompact/PostCompact hooks so nothing is ever lost.

A local-first agent memory lifecycle engine: typed memories in an append-only, integrity-chained log with dedupe, conflict detection, forgetting, compaction, verification, query, and portable export.

Olake-Fusion: A iceberg table maintenance tool managed by Olake.

Optimize rocksdb compaction

Delta Lake Optimization Project: Hands‑on lab to explore partitioning, Z‑Ordering, compaction (manual & auto), Liquid Clustering, and VACUUM using a synthetic sales dataset in Databricks. Includes a step‑by‑step notebook to measure file scans, bytes read, and query performance for each optimization.

Extractive document compression for RAG and agent pipelines. +69% vs LLMLingua, +175% vs LLMLingua2 on information recovery. Compiled wheels, try it out.

Claude Code harness that refreshes its own context instead of compacting: at a threshold it writes a verified capsule, runs /clear, and resumes in a fresh session. Memory is plain markdown in git, readable with git log. MIT, runs locally, one install (Node 18+). Codex adapter included.