Data
Plugin
openviking
Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
Plugin Signal83/100Strong
maintained99
capable46
documented100
adopted95
Install on Claude Code
> /plugin marketplace add volcengine/openviking> /plugin install openviking-memory@openviking
What's inside
Skills
8/cronSchedule reminders and recurring tasks./github-proxyGitHub ๅฝๅ
่ฎฟ้ฎๅ ้ skill๏ผไฝฟ็จ githubproxy.cc ไปฃ็ๅ ้ GitHub ไปๅบๅ
้ใๆไปถไธ่ฝฝใRaw ๆไปถ่ฎฟ้ฎ็ญๆไฝใไฝฟ็จๅบๆฏ๏ผ(1) ้่ฆ git clone GitHub ไปๅบๆถๅ ้๏ผ(2) ไธ่ฝฝ GitHub Release ๆไปถใRaw ๆไปถใArchive ๅ็ผฉๅ
ๆถๅ ้๏ผ(3) ไปปไฝ้่ฆ่ฎฟ้ฎ GitHub ่ตๆบไฝ้ๅบฆๆ
ข็ๅบๆฏ/githubInteract with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries./opencodeMonitor and manage your OpenCode tasks using helper scripts./skill-creatorCreate or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets./summarizeSummarize or extract text/transcripts from URLs, podcasts, and local files (great fallback for โtranscribe this YouTube/videoโ).
README
1OpenViking: The Context Database for AI Agents
Website ยท Live Demo ยท GitHub ยท Issues ยท Docs
๐ Join our Community
๐ฑ Lark Group ยท WeChat ยท Discord ยท X
What is OpenViking
OpenViking is an open-source context database for AI agents. It stores memories, resources, and skills as one virtual filesystem under the viking:// protocol, so an agent browses its own context with ls, tree, and find instead of querying a black-box vector store. Content is processed into three tiers โ L0 abstract, L1 overview, L2 details โ and loaded on demand. Every retrieval leaves a trajectory you can watch and debug. Full introduction: Getting started.
The OpenViking Studio playground โ a live demo you can open in the browser, no installation required.
Why OpenViking
- One filesystem for all context. Memories, resources, and skills each get a
viking://URI. Agents locate and manipulate context deterministically, like a developer working with files. โ Viking URI ยท Context types - Tiered loading cuts token spend. Every entry is processed into L0 (abstract), L1 (overview), and L2 (details) on write, then loaded only as deep as the task requires. โ Context layers
- Directory recursive retrieval. Vector search first locates the highest-scoring directory, then drills down layer by layer, so results arrive with their surrounding context intact. โ Retrieval
- Observable retrieval. Each query preserves its directory-browsing trajectory. When a result looks wrong, you can see exactly which path produced it. โ Retrieval
- Sessions become memory. After a session commits, OpenViking asynchronously extracts user preferences and agent experience into long-term memory. โ Session
How the pieces fit together: Architecture. The thinking behind the design: The Database Paradigm for Context Engineering.
viking://
โโโ resources/ # Resources: project docs, repos, web pages, etc.
โ โโโ my_project/
โ โโโ docs/
โ โ โโโ api/
โ โ โโโ tutorials/
โ โโโ src/
โโโ user/
โโโ {user_id}/
โโโ memories/
โ โโโ preferences/
โ โโโ writing_style
โ โโโ coding_habits
โโโ resources/
โ โโโ private_project/
โโโ skills/
โ โโโ search_code
โ โโโ analyze_data
โโโ peers/
โโโ web-visitor-alice/
The three loading tiers:
- L0 (Abstract): a one-sentence summary for quick relevance checks.
- L1 (Overview): core information and usage scenarios for planning.
- L2 (Details): the full original data, read only when needed.
Each directory carries its own L0/L1 layers, so relevance can be judged before any full file is read:
viking://resources/my_project/
โโโ .abstract # L0: ~100 tokens - quick relevance check
โโโ .overview # L1: ~2k tokens - structure and key points
โโโ docs/
โโโ .abstract
โโโ .overview
โโโ api/
โโโ auth.md # L2: full content, loaded on demand
โโโ endpoints.md
Proof it works
OpenViking 0.3.22 has been evaluated on long-conversation user memory (LoCoMo) and multi-turn agent tasks (tau2-bench). Full results and setup details, including knowledge-base QA, are in the benchmark report; reproduction scripts live in ./benchmark.
- User memory (LoCoMo): with OpenViking, all three agent integrations land at 80โ83% accuracy โ up from 24โ57% on their native memory โ while input tokens drop by 34.3โ91.0% and query latency by 58.45โ66.10%.
- Agent experience (tau2-bench): experience memory lifts task success by +6.87pp (retail) and +11.87pp (airline) over the same LLM without memory.
Quick start
๐ก Want to see it in action first? Try OpenViking Studio โ a live hosted instance with a context playground, semantic search, and a multi-agent hub. No installation required.
Requires Python 3.10 or higher.
pip install openviking --upgrade
openviking-server init # interactive wizard: providers, models, ov.conf
openviking-server doctor # validate setup
openviking-server # start (background: nohup openviking-server > openviking.log 2>&1 &)
init walks you through provider setup and writes ~/.openviking/ov.conf. It supports Volcengine, OpenAI, Codex OAuth, Kimi, GLM, and local Ollama โ for Ollama it can detect and install the runtime and pull models suited to your hardware. doctor checks the config file, Python version, provider connectivity, and disk space without a running server. Manual ov.conf templates, per-provider examples, environment variables, and Windows setup: Configuration guide ยท Quick start docs.
The install already includes the ov client CLI. With the server running:
ov status
ov add-resource https://github.com/volcengine/OpenViking # --wait
ov ls viking://resources/
ov tree viking://resources/volcengine -L 2
# wait some time for semantic processing if not --wait
ov find "what is openviking"
ov grep "openviking" --uri viking://resources/volcengine/OpenViking/docs/en
Next steps:
- Client configuration (
ov config), standalone CLI installs (npm / cargo), and advanced usage such as index rebuilding: CLI setup - Docker and production deployment: Deployment guide
Use it with your agent
Integrations inject OpenViking recall into your agent's context and auto-commit session memory:
Setup instructions for each agent: Agent integrations overview.
OpenViking Helper (Beta)
OpenViking Helper is a desktop console, currently in beta for macOS and Windows x64:
- Visual local agent setup: detects OpenViking CLI, Claude Code, Codex, Cursor, Trae, and OpenCode, then configures supported plugin, MCP, Hook, and CLI integrations.
- Session trace inspection: parses Claude Code, Codex, and Trae sessions to show OpenViking recall, prompt injection, MCP calls, capture, and commit events.
- Local memory and skill management: views local memory / rule files and
SKILL.mdskills, then syncs them to OpenViking.
Download:
VikingBot
VikingBot is an AI agent framework built on top of OpenViking:
pip install "openviking[bot]"
openviking-server --with-bot
ov chat # in another terminal
The official Docker image bundles VikingBot and starts it by default alongside the server and console UI. Details: VikingBot guide.
Deploy in production
For production, run OpenViking as a standalone HTTP service โ see Server deployment and the Deployment guide.
Prefer not to operate it yourself? OpenViking Personal is officially hosted and ready to use, scales far beyond local hardware with VikingDB, and includes a free trial for up to 50 files; existing open-source users can move over with the migration tool. โ openviking.ai
Managed Services
Volcano Engine (China Region)
Officially hosted on Volcano Engine, available in China.
- Product page: www.volcengine.com/product/openviking-service
- Documentation: docs.volcengine.com/docs/84313/2374478
Coming Soon
- Global hosting service on BytePlus (for regions outside China)
- Private distributed enterprise edition (deployed via activation code, supports BYOC)
Research
OpenViking open-sources a subset of the core capabilities described in the VikingMem paper:
VikingMem: A Memory Base Management System for Stateful LLM-based Applications Jiajie Fu, Junwen Chen, Mengzhao Wang, Aoxiang He, Maojia Sheng, Xiangyu Ke, Yifan Zhu, and Yunjun Gao. arXiv:2605.29640, 2026. Accepted by VLDB 2026. ๐ Read the paper on arXiv
Partner Projects
OpenViking welcomes collaboration with other open-source projects to build the context data ecosystem. Our confirmed partners include:
- deer-flow - Open-source long-horizon SuperAgent harness
- NoKV - AI native distributed file system
- loopx - Lightweight loop engineering state kernel
- Hermes Agent - The agent that grows with you
Interested in joining our partner list? Please submit an issue to our community to apply.
Community & contributing
OpenViking is still in its early stages, and there is plenty left to build.
- Docs: docs.openviking.ai ยท FAQ
- Blog: blog.openviking.ai
- Team: About us
- Chat: ๐ฑ Lark Group ยท ๐ฌ WeChat ยท ๐ฎ Discord ยท ๐ฆ X
- Contribute: bug fixes and new features are both welcome โ see CONTRIBUTING.md
Security and privacy
This project takes security seriously. For vulnerability reporting and supported versions, see SECURITY.md
License
The OpenViking project uses different licenses for different components:
File tree
1652volcengine/openvikingView on GitHub
.agents/
plugins/
marketplace.json
.clang-format
.claude-plugin/
marketplace.json
.github/
dependabot.yml
ISSUE_TEMPLATE/
bug_report.yml
config.yml
feature_request.yml
question.yml
PULL_REQUEST_TEMPLATE.md
scripts/
stage-memory-plugin-marketplace.sh
workflows/
_build.yml
_codeql.yml
_docs-deploy.yml
_docs.yml
_publish.yml
_test_full.yml
_test_lite.yml
api_test_effect.yml
api_test.yml
build-docker-image.yml
ci.yml
clawhub-dev-release.yml
controlplane-mcp-release.yml
docs-tos.yml
docs.yml
oc2ov_test.yml
opencode-plugin-release.yml
pr.yml
python-langchain-release.yml
python-sdk-release.yml
release-tos.yml
release-vikingbot-first.yml
release.yml
rust-cli.yml
schedule.yml
typescript-sdk-release.yml
.gitignore
.pr_agent.toml
benchmark/
__init__.py
.gitignore
custom/
README.md
session_contention_benchmark.py
cuvs/
PRELIMINARY_RESULTS.md
README.md
run_collection_benchmark.py
run_index_benchmark.py
run_service_concurrency_benchmark.py
summarize_collection_runs.py
summarize_index_runs.py
summarize_service_runs.py
locomo/
claudecode/
config/
ov-hooks.json
ov-ingest.conf
ov-mcp.json
ov-qa.conf
sys-prompts/
ack.txt
summary.txt
eval.py
import_to_ov.py
ingest_e2e.py
ingest.py
judge.py
README.md
run_e2e.sh
run_prompted.sh
run_sdk_iso.sh
run_sdk_noiso.sh
scripts/
auto-capture.mjs
stat_judge_result.py
hermes/
eval.py
import_e2e.py
import_to_native.py
import_to_ov.py
judge.py
README.md
run_full_eval.sh
stat_judge_result.py
locomo_bad_case_questions.csv
mem0/
delete_user.py
eval.py
ingest.py
README.md
openclaw/
eval.py
import_to_ov.py
judge.py
run_full_eval.sh
stat_judge_result.py
openviking/
import_to_ov.py
judge.py
locomo_prompts.py
README.md
run_eval.py
stat_judge_result.py
README.md
supermemory/
delete_container.py
eval.py
ingest.py
README.md
vikingbot/
import_to_ov.py
judge.py
preflight_eval_config.py
preflight_eval_runtime.py
progress_utils.py
run_eval.py
run_full_eval.sh
stat_judge_result.py
longmemeval/
openviking/
import_to_ov.py
judge.py
longmemeval_prompts.py
README.md
run_eval.py
stat_judge_result.py
RAG/
.gitignore
config/
config.yaml
financebench_config.yaml
locomo_config.yaml
qasper_config.yaml
syllabusqa_config.yaml
ov.conf.example
README_zh.md
README.md
run.py
scripts/
__init__.py
download_dataset.py
prepare_dataset.py
run_sampling.py
sample_dataset.py
src/
__init__.py
adapters/
__init__.py
base.py
financebench_adapter.py
locomo_adapter.py
qasper_adapter.py
syllabusqa_adapter.py
core/
__init__.py
judge_util.py
llm_client.py
logger.py
metrics.py
monitor.py
vector_store.py
pipeline.py
retrieval/
grep/
vikingdb_bm25/
ai_wiki.txt
effectiveness/
step1_add_resource.py
step2_quality.py
performance/
step0_prepare_data.py
step1_add_resource.py
step2_reindex.py
step3_benchmark.py
README_CN.md
README.md
skillsbench/
skill_bench_eval.py
tau2/
__init__.py
common/
__init__.py
tau2_env/
__init__.py
tau2_environment.py
tau2_tool_provider.py
llm/
.gitignore
config/
baseline.yaml
fixed_first_user_bootstrap.yaml
no_memory.yaml
scope_prompts/
generic_memory_scope.md
template_indexed_trajectory.yaml
README.md
run_full_eval.sh
scripts/
build_fixed_first_user_fixture.py
run_eval.py
run_memory_v2_eval.py
setup_tau2_repo.sh
tau2_common.py
train/
__init__.py
_rollout_helpers.py
case_loader.py
experience_loader_template/
SKILL.md
README.md
restart_vikingbot_train_eval.sh
rollout_executor_native.py
rollout_executor_vikingbot.py
rollout_executor.py
run_batch_train_eval.sh
run_service.sh
service_app.py
vikingbot/
.gitignore
README.md
run_airline_2epochs.sh
run_full_test.sh
scripts/
commit_trajectory_to_memory.py
provision_openviking_user.py
run_eval_reward.sh
run_tau2_domain.sh
vikingbot_tau2_runner.py
setup_env.sh
vectordb_perf/
__init__.py
async_utils.py
README.md
run.py
bot/
.coveragerc
.dockerignore
.github/
workflows/
release.yml
test.yml
.gitignore
bridge/
package.json
src/
index.ts
server.ts
types.d.ts
whatsapp.ts
tsconfig.json
demo/
werewolf/
Cubic_11_1.010_R.ttf
GeistPixel-Square.woff2
README.md
SOUL-god.md
SOUL-player.md
start_werewolf_demo.py
werewolf_server.py
werewolfUI.html
deploy/
docker/
build-image.sh
build-multiarch.sh
deploy_langfuse.sh
deploy.sh
docker-entrypoint.sh
Dockerfile
image_upload.example.yaml
image_upload.sh
langfuse/
docker-compose.yml
README.md
stop.sh
Dockerfile
ecs/
README.md
vke/
deploy.sh
k8s/
deployment.yaml
pvc-nas-example.yaml
pvc-tos-example.yaml
pvc-tos.yaml
README.md
vke_deploy.example.yaml
docs/
en/
concepts/
01-architecture.md
02-agent-capabilities.md
03-channels-and-gateway.md
04-openviking-integration.md
05-channel.md
zh/
concepts/
01-architecture.md
02-agent-capabilities.md
03-channels-and-gateway.md
04-openviking-integration.md
05-channel.md
design/
rfc-openviking-cli-ov-chat.md
vikingbot-feedback-observability-design.md
vikingbot-langfuse-local-setup.md
vikingbot-openviking-context-plan.md
license/
LICENSE
package-lock.json
package.json
README_CN.md
README.md
scripts/
clean_vikingbot.sh
install_local_openclaw_plugin.sh
kill_openviking_server.sh
restart_openviking_server.sh
start_vikingbot_in_ecs.sh
test_all.sh
test_restart_openviking_server.sh
tests/
conftest.py
example.py
experience_data_mini.json
test_agent_loop_outcome.py
test_channel_delivery_metadata.py
test_channel_sender_name.py
test_chat_functionality.py
test_compile.py
test_feedback_stats.py
test_gateway_startup_security.py
test_image_format.py
test_image_tool_sandbox.py
test_langfuse_outcome_metadata.py
test_minimax_provider.py
test_openapi_auth.py
test_openviking_api_key_type.py
test_outcome_evaluator.py
test_sandbox_startup.py
test_skills_metadata.py
test_subagent_skills_context.py
test_werewolf_server_security.py
vikingbot/
__init__.py
__main__.py
agent/
__init__.py
context.py
loop.py
memory.py
skills.py
subagent.py
tools/
__init__.py
base.py
compile.py
cron.py
factory.py
filesystem.py
image.py
mcp.py
message.py
ov_file.py
registry.py
shell.py
spawn.py
web.py
websearch/
__init__.py
base.py
brave.py
ddgs.py
exa.py
registry.py
tavily.py
bus/
__init__.py
events.py
queue.py
channels/
__init__.py
base.py
botchannel.py
chat.py
dingtalk.py
discord.py
email.py
feishu_test.py
feishu.py
manager.py
mochat.py
openapi_models.py
... 1600 moreShowing a partial view of a very large repo.
FAQ
openviking is a Claude Code plugin with 8 hand-picked skills for data work, indexed on Flowy. Install it with the command on its page. It includes cron, github-proxy, github. Its skills do not fire on their own yet. Request auto-invocation to have Flowy route them as you prompt. Free and open source.
- What is openviking?
- Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
- How do I install openviking?
- Run these in Claude Code: /plugin marketplace add volcengine/openviking then /plugin install openviking-memory@openviking. Then prompt normally.
- Does openviking auto-invoke its skills?
- Not yet. It is indexed on Flowy as a plain plugin. Request auto-invocation on its page and Flowy will route its skills for you as you prompt.
- Is openviking free?
- Yes. Flowy is free and open source, with nothing gated. You can read every skill in full before you install.