AI agent memory for engineering teams
team memory

Palace turns agent memory from a personal convenience into a shared engineering asset, while keeping it self-hosted.

Works with Cursor, Claude Code, Codex, and other MCP-capable clients.
Combines semantic retrieval, BM25-style search, diaries, and knowledge graph recall.
Open-source MIT core for solo developers; paid server for shared team memory.

The team problem

Every developer teaches their agent the same conventions, build quirks, and architectural decisions. Without shared memory, that context resets over and over.

The Palace approach

Palace captures durable memories with provenance and makes them searchable to future agent sessions across the team.

The buyer reason

Paid Palace Server gives teams a controlled deployment, flat pricing, seat enforcement, and operational ownership of the memory layer.