Temporal memory
Keep decisions, changes and outcomes in context.
v0.19.3 · Meet team memoryYour next session.
Your next session.
Already in context.
Give your Agent a lasting home for knowledge, decisions and plans. Keep it local, or share it with your team. Pick up where the work left off.
Open source. Local ownership. No embedded model.
Knowledge workspaceLocal + shared
Team guide
A shared place to begin
Keep the context your team needs, ready for the next person and the next task.
- Find knowledge with its sources
- Share only with the right people
- Continue work with your Agent
Context travels with the work
Sources and change history remain traceable.
Built around your memory
- Local firstYour memory stays with you.
- Agent operatedOpen tools. Your choice of Agent.
- Shared by permissionDecide who can access each space.
- Open sourceRun it on your own infrastructure.
A simple rhythm
From useful information
to lasting knowledge.
Your Agent does the work. LWC keeps the continuity.
- 01
Keep what matters
Capture sources, decisions and useful context as you work.
- 02
Make knowledge usable
Connect pages, evidence and plans in a searchable knowledge space.
- 03
Continue with context
Your Agent can pick up the next task with the knowledge already in place.
One place for continuity
More than remembering a note.
Keep knowledge, unfinished work and learning connected.
A knowledge space you can build on
Searchable pages, linked sources and a record of what changed. Ready for people to read and Agents to use.
Durable Todo
Keep unfinished work visible across sessions.
Current Plan
Follow clear steps and retain progress and evidence.
Tutor
Continue a learning path from where you left off.
Book
Read in source order, with traceable notes and citations.
Practice
Build on exercises, feedback and scheduled review.
Built for your team
Share knowledge.
Keep control.
The same local-first foundation, with shared spaces and clear permissions. Host it where your team needs it.
Read the team guideA straightforward way in
Generated personal keys, space selection, and member setup with initial access.
Sync that follows the work
LWC moves authorized changes automatically. Your external Agent resolves semantic conflicts with the evidence preserved.
Recover with confidence
Explicit access policies, auditable recovery and backups that preserve current permissions.
Get started
Let your Agent
take it from here.
Copy the setup instructions and send them to your Agent. It can install, configure and verify LWC for your environment.
Prefer to install manually?Ready-to-use setup instructionsOne copy
- Copy the instructions below.
- Send them to your preferred Agent.
- Let it configure and verify the installation.
Read the full instructions
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.Keep exploring
Everything you need, within reach.
DocumentationGuides and referenceDownloadsRelease notes and packagesSource codeExplore and contributeFeedbackReport issues and ideas
Agent behavior and boundaries
No-code learning never prompts for CodeGraph. Bound Tutor turns stay silent and display only after commit. Before a meaningful batch, phase change, or visible wait, the Agent gives one outcome-level sentence.