OpenViking Adapter for Openclaw

A high-efficiency memory adapter for OpenClaw that implements ByteDance's OpenViking architecture to optimize context window usage.

yyu812707-wq
v1.0.0
Mar 5, 2026
0
869
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install openviking-adapter

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install openviking-adapter using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is OpenViking Adapter?

The OpenViking Adapter is a sophisticated implementation of ByteDance's open-source AI memory system designed specifically for the Openclaw Skills ecosystem. It addresses the common challenge of information overload and high token costs in complex agent workflows. By utilizing a three-layer memory architecture, it ensures that agents retain critical context while filtering out unnecessary data, leading to a significant 83% reduction in token consumption and boosting task completion rates from 35% to 52%.

OpenViking Adapter Use Cases

  • Reducing operational costs for long-running AI conversations by optimizing token usage.
  • Improving the accuracy of AI agents in complex environments through structured memory recall.
  • Managing large datasets or long histories without exceeding LLM context limits.
  • Implementing a consistent 'identity' for agents across multiple sessions using the L0 Soul layer.

How OpenViking Adapter Works

  1. The adapter performs an initial token audit to identify current memory inefficiencies.
  2. It generates an L0 Soul layer, a compact 100-token summary of the agent's core identity and mission.
  3. An L1 Overview layer is constructed to capture up to 2000 tokens of high-level interaction history.
  4. For specific queries, the system triggers the L2 Detail layer to perform an on-demand search for relevant granular memories.
  5. The optimization engine dynamically loads only the most relevant layers into the context window for each request.

OpenViking Adapter Setup

Install the adapter through the command line:

clawhub install openviking-adapter

Analyze your current memory footprint:

openclaw tools call openviking-adapter analyze_token_usage

Run the optimization process:

openclaw tools call openviking-adapter optimize_memory_loading

OpenViking Adapter Data Schema & Taxonomy

The adapter organizes memory into three distinct tiers for maximum efficiency:

Tier Identifier Token Limit Description
L0 Soul Layer ~100 Core identity and personality summary
L1 Overview Layer ~2000 Summarized historical context and key events
L2 Detail Layer Search-based Granular memory fragments retrieved as needed

OpenViking Adapter Advanced Features

  • Intelligent token usage analysis to monitor cost-performance ratios within Openclaw Skills.
  • Automated generation of core identity summaries to maintain agent consistency.
  • On-demand semantic search for L2 detailed memories to minimize context bloat.
  • Full optimization workflow that automates the transition between memory layers based on task complexity.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*