OpenCode MemPalace Plugin

OpenCode memory layer with private retrieval and safe autosave.

Русская версия

Why it matters

Keep OpenCode context-aware without extra prompts.

• Finds relevant memory automatically before normal replies
• Saves durable knowledge quietly in the background
• Keeps writes safe through one controlled tool
• Protects sensitive data before anything is stored

Quick start

Install MemPalace:

pip install mempalace
mempalace init <dir>
mempalace mine <dir>

Add the plugin to opencode.json:

{
  "plugin": ["@rvboris/opencode-mempalace"]
}

That is enough to enable memory search, autosave, and mempalace_memory.

What you get

• hidden memory lookup before answers
• autosave on session lifecycle events
• separate user and project memory
• one safe memory tool for the model
• local Python bridge to MemPalace

The plugin does not require the MemPalace MCP server.

The one tool: mempalace_memory

This is the only tool the model needs.

Modes

• save — store a durable preference, fact, or decision
• search — find relevant memory by query
• kg_add — add a structured fact to the knowledge graph
• diary_write — save a short work note or daily log
• mine_messages — internal autosave mode used by the plugin

Examples

Save a user preference:

mempalace_memory
  mode: save
  scope: user
  room: preferences
  content: Prefers concise responses and numbered steps.

Save a project decision:

mempalace_memory
  mode: save
  scope: project
  room: decisions
  content: Use Bun for builds and tests.

Search memory:

mempalace_memory
  mode: search
  scope: project
  room: workflow
  query: build command
  limit: 3

Add a graph fact:

mempalace_memory
  mode: kg_add
  subject: my-repo
  predicate: uses
  object: bun

Memory areas

User memory

• preferences
• workflow
• communication

Use it for stable cross-project habits and preferences.

Project memory

• architecture
• workflow
• decisions
• bugs
• setup

Use it for repository-specific knowledge.

Configuration

Optional config file: ~/.config/opencode/mempalace.jsonc

{
  "autosaveEnabled": true,
  "retrievalEnabled": true,
  "keywordSaveEnabled": true,
  "maxInjectedItems": 6,
  "retrievalQueryLimit": 5,
  "privacyRedactionEnabled": true
}

Useful environment variables:

• MEMPALACE_AUTOSAVE_ENABLED
• MEMPALACE_RETRIEVAL_ENABLED
• MEMPALACE_KEYWORD_SAVE_ENABLED
• MEMPALACE_PRIVACY_REDACTION_ENABLED
• MEMPALACE_AUTOSAVE_LOG_FILE
• MEMPALACE_ADAPTER_PYTHON

Privacy

• respects <private>...</private> blocks
• redacts common secrets before writes
• skips fully private content

Project docs

• Release history: CHANGELOG.md
• Changelog rules: CONTRIBUTING.md#changelog

Local development

Load from source:

{
  "$schema": "https://opencode.ai/config.json",
  "plugin": [
    "file:///ABSOLUTE/PATH/TO/mempalace-autosave/plugin/index.ts"
  ]
}

Build:

npm run build

Debug logs:

opencode --log-level DEBUG

File log: ~/.mempalace/opencode_autosave.log

Links

• OpenCode: https://opencode.ai
• MemPalace: https://github.com/milla-jovovich/mempalace