Proactive memory for AI Agents

Memory that survives the session.

LWC gives AI Agents a durable, source-grounded Wiki they can recall, maintain, and improve across coding, research, operations, and knowledge work.

Install with your Agent Explore the Wiki

Trust by design

How it works

From evidence to useful recall.

LWC makes memory a maintained project asset: grounded in sources, shaped into a Wiki, and loaded only when the current task needs it.

  1. 01 / Add evidence

    Bring in what matters.

    Record pages, ingest sources, or convert documents while preserving where knowledge came from.

  2. 02 / Maintain the Wiki

    Keep memory coherent.

    Search, lint, tag, revise, and connect knowledge through audited CLI workflows.

  3. 03 / Recall with context

    Start informed.

    Agents load bounded memory, strong tags, and graph context at the moment they become useful.

Capability layers

More than a folder of notes.

Use the durable Wiki on its own, then enable graph capabilities where relationships and code structure justify them.

Core layer

Persistent Wiki memory

Source-backed pages, retrieval, context loading, tags, changesets, checkpoints, and structured JSON contracts form the durable foundation.

  • Relationship layer

    Document knowledge graph

    Project pages, sources, links, provenance, and semantic relations become a verifiable physical graph.

  • Discovery layer

    Word graph

    Terms reveal bounded, high-signal connections across documents without loading the entire vocabulary at once.

  • Optional code layer

    CodeGraph

    Explore symbols, callers, files, and change impact through LWC when code structure is part of the task.

Agent installation

Ask your Agent to set it up.

Copy the complete prompt. It tells your Agent to use LWC's official installer, preserve existing configuration, verify every integration surface, and ask before enabling project memory or graphs.

AgentTarget examplesClaude Code, Codex, Cursor, OpenCode, Gemini CLI, Kiro, Hermes, Antigravity, and Pi. Examples are not exhaustive.

Prefer manual installation? Read the official options →
Complete setup prompt 
Configure LWC completely for this user. Perform and verify the work; do not
merely describe commands for me to run.

Source of truth:
- https://github.com/JanYork/llm-wiki-cli
- https://github.com/JanYork/llm-wiki-cli/tree/main/skills/using-lwc

Requirements:
1. Read this README, `SECURITY.md`, and `skills/using-lwc/SKILL.md`. Install the
   official checksum-verified release if `lwc` is not globally callable; never
   prefix routine commands with a private binary path or `LWC_PROJECT_ROOT`.
2. Run `lwc --version`, initialize global memory once with
   `lwc --scope global init` when missing, then run `lwc agent install --yes`.
   This command detects installed supported Agents and safely installs their
   MCP, Skill, Hook and Instructions using official locations. Do not recreate
   that logic manually or install a native package for the same Agent as well.
3. Inspect `lwc agent status --target all --location global`. Restart affected
   Agents and complete their normal Hook trust review where required. Do not
   initialize a project Wiki or either graph without explicit project consent.
4. If the current runtime is not one of LWC's registered AgentTargets, use its
   official user-level conventions to install the canonical `using-lwc` Skill,
   an additive instruction block, `lwc serve --mcp`, and a bounded session Hook
   only where those surfaces are officially supported. Preserve existing
   configuration, remain idempotent, and report unsupported surfaces instead of
   inventing paths or keys.

Finish with the LWC version, detected and configured Targets, status results,
files changed, unsupported surfaces, and any restart or trust action remaining.

Project resources