企业官网建设流程全解析

uadmin@UD26:~$ openviking-server No OpenViking configuration found. Run interactive setup now? [Y/n]: Y ╭─────────────────────────────────────────────────────────╮ │ OpenViking Setup │ │ Context database for AI agents — data in, context out │ ╰─────────────────────────────────────────────────────────╯ Data will be stored under /home/uadmin/.openviking/data unless you edit ov.conf later. Choose setup mode: ↑/↓ move · enter select ❯ Step-by-step setup ───────────────────────────────────────────────────────────────────────────────── Step 1/2 · Embedding — powers semantic search ───────────────────────────────────────────────────────────────────────────────── Embedding setup: ↑/↓ move · enter select · ← back ❯ Cloud API Embedding provider: ↑/↓ move · enter select · ← back ❯ VolcEngine (火山引擎) Embedding configuration API Key: ark-898***********************************352a Model [doubao-embedding-vision-251215]: Dimension: 1024 (auto-filled for doubao-embedding-vision-251215) ✓ Embedding: volcengine · doubao-embedding-vision-251215 (1024d) ───────────────────────────────────────────────────────────────────────────── Step 2/2 · VLM — parses documents & extracts memories ───────────────────────────────────────────────────────────────────────────── VLM provider: ↑/↓ move · enter select · ← back ❯ VolcEngine (火山引擎) VolcEngine VLM configuration Reuse the embedding API key for the VLM? [Y/n]: Y Model [doubao-seed-2-0-code-preview-260215]: Doubao-Seed-2.0-lite-260428 ✓ VLM: volcengine · Doubao-Seed-2.0-lite-260428 ─────────────────────────────────────────────────────────────────────────────────────────── Query planner (optional) ──────────────────────────────────────────────────────────────────────────────────────────── A small local model that plans retrieval before search — skips lookups for small talk and emits focused queries otherwise, saving tokens. Enable a lightweight local query planner via Ollama? [y/N]: y Checking Ollama... not installed Install Ollama now? [Y/n]: Y >>> Installing ollama to /usr/local [sudo] password for uadmin: ERROR: This version requires zstd for extraction. Please install zstd and try again: - Debian/Ubuntu: sudo apt-get install zstd - RHEL/CentOS/Fedora: sudo dnf install zstd - Arch: sudo pacman -S zstd Installation failed. Try manually: https://ollama.com/download Continue without Ollama? (config will be written but the model won't be pulled) [y/N]: y Query planner model: ↑/↓ move · enter select ❯ ov_intent_analysis_sft v7_q8 ───────────────────────────────────────────────────────────────────────────────────────────────── Server & auth ───────────────────────────────────────────────────────────────────────────────────────────────── Local: only this machine can reach the server (no auth, dev mode). Remote: bind to 0.0.0.0 for Docker / LAN access — requires API-key auth. Bind server host to: ↑/↓ move · enter select ❯ Local (127.0.0.1) Port [1933]: ──────────────────────────────────────────────────────────────────────────────────────────────────── Summary ──────────────────────────────────────────────────────────────────────────────────────────────────── Embedding: volcengine · doubao-embedding-vision-251215 (1024d) VLM: volcengine · Doubao-Seed-2.0-lite-260428 Query planner: litellm · ollama/guoxuter/ov_intent_analysis_sft:v7_q8 Server: 127.0.0.1:1933 · auth dev (no auth) Workspace: configured (hidden) Config: default config location Save configuration? [Y/n]: Y OK Configuration written to the default config location Next steps: Start the server: openviking-server Validate setup: openviking-server doctor Validate the setup now? (runs `openviking-server doctor`) [Y/n]: Y OpenViking Doctor Config: PASS /home/uadmin/.openviking/ov.conf Python: PASS 3.11.15 (>= 3.10 required) Native Engine: PASS variant=x86_avx2 AGFS: PASS AGFS SDK 0.1.7 Embedding: FAIL volcengine/doubao-embedding-vision-251215 api_base=https://ark.cn-beijing.volces.com/api/v3 dimension=1024 (probe failed: Volcengine embedding failed: Error code: 404 - {'error': {'code': 'ModelNotOpen', 'message': 'Your account 2102587980 has not activated the model doubao-embedding-vision-251215. Please activate the model service in the Ark Console. Request id: 0217847604875489209c58391d56dc4f2d7714e68ed22bf72d1c1', 'param': '', 'type': 'Not Found'}}, request_id: ToB-direct,OpenViking_Service,openviking-service_cn-beijing) Fix: Check the embedding provider service, model name, API key, and api_base VLM: PASS volcengine/Doubao-Seed-2.0-lite-260428 Ollama: PASS running at localhost:11434 VikingBot: PASS VikingBot aligned with dev OpenViking auth Disk: PASS 948.4 GB free in /home/uadmin/.openviking/data 1 check(s) failed. See above for fix suggestions. Start the server now? [y/N]: y Ollama is running at localhost:11434 OpenViking HTTP Server is running on 127.0.0.1:1933 2026-07-23 08:06:42.178 | INFO | openviking.telemetry.tracer:init_tracer_from_server_config:233 - [TRACER] disabled in server.observability.traces [07/23/26 08:06:42] INFO StreamableHTTP session manager started streamable_http_manager.py:131 [07/23/26 08:07:14] WARNING Run time of job "PersistCollection._register_index_manage_job (trigger: date[2026-07-23 08:07:12 CST], next run at: 2026-07-23 08:07:12 CST)" was missed by base.py:126 0:00:02.720525

环境
Win11，WSL2，Ubuntu24.04

一、相关文档

记忆提供程序

Memory Providers

openviking简介

openviking快速开始

二、前置安装

虚拟环境

openviking - Hermes Agent

隔离 Python 环境​

Hermes 通过 HTTP 连接 OpenViking，因此无需把 OpenViking 安装到 Hermes 的 Python 环境中。请在独立的虚拟环境或容器中运行 OpenViking 服务。不要在 已有 Hermes 的环境中使用--force-reinstall安装或升级 OpenViking：Hermes 版本可能会固定与 OpenViking 已支持、已修复安全问题的版本不同的依赖。如果确实要将 两个应用放在同一环境中，请在同一次依赖求解中安装它们，并在启动任一服务前运行python -m pip check。

安装虚拟环境命令

venv
python3 -m venv ~/.venvs/openviking
source ~/.venvs/openviking/bin/activate
pip install openviking

安装pipx
sudo apt install pipx && pipx ensurepath
pip show pipx

pipx install openviking

安装uv（建议）
curl -LsSf https://astral.sh/uv/install.sh | sh
uv --version

uv tool install openviking （推荐）全局 CLI 工具安装，相当于pipx，
uv pip install openviking 项目级依赖安装（安装到当前激活的项目虚拟环境中）用 uv venv + uv pip 替代 venv+pip

conda，anaconda

...

三、安装openviking

uadmin@UD26:~$ uv tool install openviking

...

Installed 4 executables: openviking, openviking-server, ov, vikingbot

四、配置openviking

配置准备

#初始化（配置文件）

openviking-server init

#配置检查
openviking-server doctor

启动服务

openviking-server（我直接执行的这个）

没初始化的化，先调用初始化

火山引擎（火山方舟）为例

openviking-server操作选择

uadmin@UD26:~$ openviking-server No OpenViking configuration found. Run interactive setup now? [Y/n]: Y ╭─────────────────────────────────────────────────────────╮ │ OpenViking Setup │ │ Context database for AI agents — data in, context out │ ╰─────────────────────────────────────────────────────────╯ Data will be stored under /home/uadmin/.openviking/data unless you edit ov.conf later. Choose setup mode: ↑/↓ move · enter select ❯ Step-by-step setup (pick embedding & VLM separately — cloud, local, or mixed) Recommended local setup (all-Ollama, sized to your RAM, one confirm) Manual (edit ov.conf directly) ---------------------------------------------------------------------------------------- ❯ Step-by-step setup ───────────────────────────────────────────────────────────────────────────────── Step 1/2 · Embedding — powers semantic search ───────────────────────────────────────────────────────────────────────────────── Embedding setup: ↑/↓ move · enter select · ← back ❯ Cloud API (VolcEngine, BytePlus, OpenAI) Local via Ollama (no API key, runs on this machine) Lightweight CPU embedding (llama.cpp, ~24 MB, no Ollama needed) ---------------------------------------------------------------------------------------- Embedding setup: ↑/↓ move · enter select · ← back ❯ Cloud API Embedding provider: ↑/↓ move · enter select · ← back ❯ VolcEngine (火山引擎) BytePlus OpenAI Other (manual) ---------------------------------------------------------------------------------------- ❯ VolcEngine (火山引擎) Embedding configuration API Key: ---------------------------------------------------------------------------------------- Embedding configuration API Key: ark-898***********************************352a Model [doubao-embedding-vision-251215]: ---------------------------------------------------------------------------------------- Dimension: 1024 (auto-filled for doubao-embedding-vision-251215) ✓ Embedding: volcengine · doubao-embedding-vision-251215 (1024d) ───────────────────────────────────────────────────────────────────────────── Step 2/2 · VLM — parses documents & extracts memories ───────────────────────────────────────────────────────────────────────────── VLM provider: ↑/↓ move · enter select · ← back ❯ VolcEngine (火山引擎) (API) BytePlus (API) OpenAI (API) OpenAI Codex (Subscription) Kimi (Subscription API Key) GLM (Subscription API Key) Custom (OpenAI-compatible) (Any OpenAI-compatible endpoint) Local via Ollama (no API key, runs on this machine) Skip for now (embedding only — add a VLM later) ---------------------------------------------------------------------------------------- ❯ VolcEngine (火山引擎) VolcEngine VLM configuration Reuse the embedding API key for the VLM? [Y/n]: ---------------------------------------------------------------------------------------- Reuse the embedding API key for the VLM? [Y/n]: Y Model [doubao-seed-2-0-code-preview-260215]: ---------------------------------------------------------------------------------------- Model [doubao-seed-2-0-code-preview-260215]: doubao-seed-2-0-lite-260428 ✓ VLM: volcengine · doubao-seed-2-0-lite-260428 ─────────────────────────────────────────────────────────────────────────────────────────── Query planner (optional) ──────────────────────────────────────────────────────────────────────────────────────────── A small local model that plans retrieval before search — skips lookups for small talk and emits focused queries otherwise, saving tokens. Enable a lightweight local query planner via Ollama? [y/N]: 轻量查询规划器，默认使用意图分析模型，ov_intent_analysis_sft:v7_q8 ---------------------------------------------------------------------------------------- Enable a lightweight local query planner via Ollama? [y/N]: y Checking Ollama... not installed # 可自行通过ollama list验证。 Install Ollama now? [Y/n]: 可以先在WSL里安Ollama，后续可以改成使用WSL外Windows主机上的Ollama ---------------------------------------------------------------------------------------- Install Ollama now? [Y/n]: Y >>> Installing ollama to /usr/local [sudo] password for uadmin: ---------------------------------------------------------------------------------------- [sudo] password for uadmin: ERROR: This version requires zstd for extraction. Please install zstd and try again: - Debian/Ubuntu: sudo apt-get install zstd - RHEL/CentOS/Fedora: sudo dnf install zstd - Arch: sudo pacman -S zstd Installation failed. Try manually: https://ollama.com/download Continue without Ollama? (config will be written but the model won't be pulled) [y/N]: 安装失败也没空关系，他会把配置配上，但不会拉模型，需要手动pull下。 ---------------------------------------------------------------------------------------- Continue without Ollama? (config will be written but the model won't be pulled) [y/N]: y Query planner model: ↑/↓ move · enter select ❯ ov_intent_analysis_sft v7_q8 ~0.8B, recommended ov_intent_analysis_sft v4_q8 ~0.8B ---------------------------------------------------------------------------------------- ❯ ov_intent_analysis_sft v7_q8 ───────────────────────────────────────────────────────────────────────────────────────────────── Server & auth ───────────────────────────────────────────────────────────────────────────────────────────────── Local: only this machine can reach the server (no auth, dev mode). Remote: bind to 0.0.0.0 for Docker / LAN access — requires API-key auth. Bind server host to: ↑/↓ move · enter select ❯ Local (127.0.0.1) (default, safer — no auth needed) Remote (0.0.0.0) (Docker / remote access — root API key required) ---------------------------------------------------------------------------------------- ❯ Local (127.0.0.1) Port [1933]: 日后需要验证时在改 ---------------------------------------------------------------------------------------- Port [1933]: ──────────────────────────────────────────────────────────────────────────────────────────────────── Summary ──────────────────────────────────────────────────────────────────────────────────────────────────── Embedding: volcengine · doubao-embedding-vision-251215 (1024d) VLM: volcengine · doubao-seed-2-0-lite-260428 Query planner: litellm · ollama/guoxuter/ov_intent_analysis_sft:v7_q8 Server: 127.0.0.1:1933 · auth dev (no auth) Workspace: configured (hidden) Config: default config location Save configuration? [Y/n]: ---------------------------------------------------------------------------------------- Save configuration? [Y/n]: Y OK Configuration written to the default config location Next steps: Start the server: openviking-server Validate setup: openviking-server doctor Validate the setup now? (runs `openviking-server doctor`) [Y/n]: 这时生成了配置文件：~/.openviking/ov.conf ---------------------------------------------------------------------------------------- Validate the setup now? (runs `openviking-server doctor`) [Y/n]: Y OpenViking Doctor Config: PASS /home/uadmin/.openviking/ov.conf Python: PASS 3.11.15 (>= 3.10 required) Native Engine: PASS variant=x86_avx2 AGFS: PASS AGFS SDK 0.1.7 Embedding: FAIL volcengine/doubao-embedding-vision-251215 api_base=https://ark.cn-beijing.volces.com/api/v3 dimension=1024 (probe failed: Volcengine embedding failed: Error code: 404 - {'error': {'code': 'ModelNotOpen', 'message': 'Your account 2102587980 has not activated the model doubao-embedding-vision-251215. Please activate the model service in the Ark Console. Request id: 0217847604875489209c58391d56dc4f2d7714e68ed22bf72d1c1', 'param': '', 'type': 'Not Found'}}, request_id: ToB-direct,OpenViking_Service,openviking-service_cn-beijing) Fix: Check the embedding provider service, model name, API key, and api_base VLM: PASS volcengine/doubao-seed-2-0-lite-260428 Ollama: PASS running at localhost:11434 VikingBot: PASS VikingBot aligned with dev OpenViking auth Disk: PASS 948.4 GB free in /home/uadmin/.openviking/data 1 check(s) failed. See above for fix suggestions. Start the server now? [y/N]: 这里建议选N，手动启动，因为会输出日志。 ---------------------------------------------------------------------------------------- 可以在Windows主机上拉模型 ollama pull guoxuter/ov_intent_analysis_sft:v7_q8 去火山方舟控制台，左侧菜单开通管理，选择语言模型（vlm使用）、向量模型（embedding使用），开通服务。 （修改vml模型：Doubao-Seed-2.0-pro-260215，Model ID为：doubao-seed-2-0-lite-260428） 开新窗口检查，openviking-server doctor ---------------------------------------------------------------------------------------- Start the server now? [y/N]: y Ollama is running at localhost:11434 OpenViking HTTP Server is running on 127.0.0.1:1933 2026-07-23 08:06:42.178 | INFO | openviking.telemetry.tracer:init_tracer_from_server_config:233 - [TRACER] disabled in server.observability.traces [07/23/26 08:06:42] INFO StreamableHTTP session manager started ----------------------------------------------------------------------------------------

有钱可选好点的vlm模型，测试可以先选简单的。

openviking-server doctor

uadmin@UD26:~$ openviking-server doctor OpenViking Doctor Config: PASS /home/uadmin/.openviking/ov.conf Python: PASS 3.11.15 (>= 3.10 required) Native Engine: PASS variant=x86_avx2 AGFS: PASS AGFS SDK 0.1.7 Embedding: PASS volcengine/doubao-embedding-vision-251215 api_base=https://ark.cn-beijing.volces.com/api/v3 dimension=1024 probe ok (dimension=1024) VLM: PASS volcengine/doubao-seed-2-0-lite-260428 Ollama: PASS running at localhost:11434 VikingBot: PASS VikingBot aligned with dev OpenViking auth Disk: PASS 948.4 GB free in /home/uadmin/.openviking/data All checks passed.

openviking-server操作日志

uadmin@UD26:~$ openviking-server No OpenViking configuration found. Run interactive setup now? [Y/n]: Y ╭─────────────────────────────────────────────────────────╮ │ OpenViking Setup │ │ Context database for AI agents — data in, context out │ ╰─────────────────────────────────────────────────────────╯ Data will be stored under /home/uadmin/.openviking/data unless you edit ov.conf later. Choose setup mode: ↑/↓ move · enter select ❯ Step-by-step setup ───────────────────────────────────────────────────────────────────────────────── Step 1/2 · Embedding — powers semantic search ───────────────────────────────────────────────────────────────────────────────── Embedding setup: ↑/↓ move · enter select · ← back ❯ Cloud API Embedding provider: ↑/↓ move · enter select · ← back ❯ VolcEngine (火山引擎) Embedding configuration API Key: ark-898***********************************352a Model [doubao-embedding-vision-251215]: Dimension: 1024 (auto-filled for doubao-embedding-vision-251215) ✓ Embedding: volcengine · doubao-embedding-vision-251215 (1024d) ───────────────────────────────────────────────────────────────────────────── Step 2/2 · VLM — parses documents & extracts memories ───────────────────────────────────────────────────────────────────────────── VLM provider: ↑/↓ move · enter select · ← back ❯ VolcEngine (火山引擎) VolcEngine VLM configuration Reuse the embedding API key for the VLM? [Y/n]: Y Model [doubao-seed-2-0-code-preview-260215]: doubao-seed-2-0-lite-260428 ✓ VLM: volcengine · doubao-seed-2-0-lite-260428 ─────────────────────────────────────────────────────────────────────────────────────────── Query planner (optional) ──────────────────────────────────────────────────────────────────────────────────────────── A small local model that plans retrieval before search — skips lookups for small talk and emits focused queries otherwise, saving tokens. Enable a lightweight local query planner via Ollama? [y/N]: y Checking Ollama... not installed Install Ollama now? [Y/n]: Y >>> Installing ollama to /usr/local [sudo] password for uadmin: ERROR: This version requires zstd for extraction. Please install zstd and try again: - Debian/Ubuntu: sudo apt-get install zstd - RHEL/CentOS/Fedora: sudo dnf install zstd - Arch: sudo pacman -S zstd Installation failed. Try manually: https://ollama.com/download Continue without Ollama? (config will be written but the model won't be pulled) [y/N]: y Query planner model: ↑/↓ move · enter select ❯ ov_intent_analysis_sft v7_q8 ───────────────────────────────────────────────────────────────────────────────────────────────── Server & auth ───────────────────────────────────────────────────────────────────────────────────────────────── Local: only this machine can reach the server (no auth, dev mode). Remote: bind to 0.0.0.0 for Docker / LAN access — requires API-key auth. Bind server host to: ↑/↓ move · enter select ❯ Local (127.0.0.1) Port [1933]: ──────────────────────────────────────────────────────────────────────────────────────────────────── Summary ──────────────────────────────────────────────────────────────────────────────────────────────────── Embedding: volcengine · doubao-embedding-vision-251215 (1024d) VLM: volcengine · doubao-seed-2-0-lite-260428 Query planner: litellm · ollama/guoxuter/ov_intent_analysis_sft:v7_q8 Server: 127.0.0.1:1933 · auth dev (no auth) Workspace: configured (hidden) Config: default config location Save configuration? [Y/n]: Y OK Configuration written to the default config location Next steps: Start the server: openviking-server Validate setup: openviking-server doctor Validate the setup now? (runs `openviking-server doctor`) [Y/n]: Y OpenViking Doctor Config: PASS /home/uadmin/.openviking/ov.conf Python: PASS 3.11.15 (>= 3.10 required) Native Engine: PASS variant=x86_avx2 AGFS: PASS AGFS SDK 0.1.7 Embedding: FAIL volcengine/doubao-embedding-vision-251215 api_base=https://ark.cn-beijing.volces.com/api/v3 dimension=1024 (probe failed: Volcengine embedding failed: Error code: 404 - {'error': {'code': 'ModelNotOpen', 'message': 'Your account 2102587980 has not activated the model doubao-embedding-vision-251215. Please activate the model service in the Ark Console. Request id: 0217847604875489209c58391d56dc4f2d7714e68ed22bf72d1c1', 'param': '', 'type': 'Not Found'}}, request_id: ToB-direct,OpenViking_Service,openviking-service_cn-beijing) Fix: Check the embedding provider service, model name, API key, and api_base VLM: PASS volcengine/doubao-seed-2-0-lite-260428 Ollama: PASS running at localhost:11434 VikingBot: PASS VikingBot aligned with dev OpenViking auth Disk: PASS 948.4 GB free in /home/uadmin/.openviking/data 1 check(s) failed. See above for fix suggestions. Start the server now? [y/N]: y Ollama is running at localhost:11434 OpenViking HTTP Server is running on 127.0.0.1:1933 2026-07-23 08:06:42.178 | INFO | openviking.telemetry.tracer:init_tracer_from_server_config:233 - [TRACER] disabled in server.observability.traces [07/23/26 08:06:42] INFO StreamableHTTP session manager started streamable_http_manager.py:131 [07/23/26 08:07:14] WARNING Run time of job "PersistCollection._register_index_manage_job (trigger: date[2026-07-23 08:07:12 CST], next run at: 2026-07-23 08:07:12 CST)" was missed by base.py:126 0:00:02.720525