vibe

VIBE AI coding sandbox config template, integrating Code-Server, JupyterLab, Claude Code and JuiceFS web development environment

The vibe config template provides a ready-to-use AI coding sandbox, integrating Code-Server (Web VS Code), JupyterLab, Claude Code CLI, JuiceFS distributed filesystem, and feature-rich PostgreSQL database.

Overview

Config Name: vibe
Node Count: Single node
Description: VIBE AI coding sandbox with Code-Server + JupyterLab + Claude Code + JuiceFS + PostgreSQL
OS Distro: el8, el9, el10, d12, d13, u22, u24
OS Arch: x86_64, aarch64
Related: meta

Usage:

./configure -c vibe [-i <primary_ip>]

Content

Source: pigsty/conf/vibe.yml

---
#==============================================================#
# File      :   vibe.yml
# Desc      :   Pigsty ai vibe coding sandbox
# Ctime     :   2026-01-19
# Mtime     :   2026-01-24
# Docs      :   https://pigsty.io/docs/conf/vibe
# License   :   Apache-2.0 @ https://pigsty.io/docs/about/license/
# Copyright :   2018-2026  Ruohang Feng / Vonng ([email protected])
#==============================================================#

# VIBE CODING SANDBOX
# PostgreSQL with related extensions
# Code-Server, Jupyter, Claude Code
#
# Usage:
#   curl https://repo.pigsty.io/get | bash
#   ./configure -c vibe
#   ./deploy.yml
#   ./juice.yml     # pgfs: juicefs on pgsql, mount on /fs
#   ./vibe.yml      # code-server, jupyter, and claude-code

all:
  children:
    infra: { hosts: { 10.10.10.10: { infra_seq: 1 }} ,vars: { repo_enabled: false }}
    etcd:  { hosts: { 10.10.10.10: { etcd_seq: 1  }} ,vars: { etcd_cluster: etcd  }}
    pgsql: { hosts: { 10.10.10.10: { pg_seq: 1, pg_role: primary } } ,vars: { pg_cluster: pgsql }}

    # optional modules
    #minio: { hosts: { 10.10.10.10: { minio_seq: 1 }} ,vars: { minio_cluster: minio }}
    #redis-ms:
    #  hosts: { 10.10.10.10: { redis_node: 1 , redis_instances: { 6379: { }, 6380: { replica_of: '10.10.10.10 6379' } } } }
    #  vars: { redis_cluster: redis-ms ,redis_password: 'redis.ms' ,redis_max_memory: 64MB }

  vars:
    #----------------------------------------------#
    # INFRA: https://pigsty.io/docs/infra
    #----------------------------------------------#
    version: v4.0.0                     # pigsty version string
    admin_ip: 10.10.10.10               # admin node ip address
    region: default                     # upstream mirror region: default,china,europe
    infra_portal:                       # infra services exposed via portal
      home : { domain: i.pigsty }       # default domain name
    dns_enabled: false                # disable dns service
    vtrace_enabled: false               # enable vtrace extension
    #blackbox_enabled: false             # disable blackbox exporter
    #alertmanager_enabled: false         # disable alertmanager
    infra_extra_services:               # home page navigation entries
      - { name: Code Server  ,url: '/code'             ,desc: 'VS Code Server'       ,icon: 'code'     }
      - { name: Jupyter      ,url: '/jupyter'          ,desc: 'Jupyter Notebook'     ,icon: 'jupyter'  }
      - { name: Claude Code  ,url: '/ui/d/claude-code' ,desc: 'Claude Observability' ,icon: 'claude'   }

    #----------------------------------------------#
    # NODE: https://pigsty.io/docs/node
    #----------------------------------------------#
    nodename_overwrite: false           # do not overwrite node hostname on single node mode
    node_tune: oltp                     # node tuning specs: oltp,olap,tiny,crit
    node_dns_method: none               # do not setup dns
    node_repo_modules: node,infra,pgsql # add these repos directly to the singleton node
    node_packages: [ openssh-server, juicefs, restic, rclone, uv, opencode, claude, code-server, golang, nodejs, asciinema, genai-toolbox, postgrest ]
    docker_enabled: true                # enable docker service
    #docker_registry_mirrors: ["https://docker.1panel.live","https://docker.1ms.run","https://docker.xuanyuan.me","https://registry-1.docker.io"]

    #----------------------------------------------#
    # PGSQL: https://pigsty.io/docs/pgsql
    #----------------------------------------------#
    pg_version: 18                      # Default PostgreSQL Major Version is 18
    pg_conf: oltp.yml                   # pgsql tuning specs: {oltp,olap,tiny,crit}.yml
    pg_packages: [ pgsql-main, patroni, pgbackrest, pg_exporter, pgbackrest_exporter ]
    pg_extensions:
      - postgis timescaledb pg_cron pgvector vchord pgvectorscale pg_search pg_textsearch vchord_bm25
      - pg_duckdb pg_mooncake pg_clickhouse pg_parquet pg_tle pljs plprql pg_stat_monitor pg_wait_sampling
      - pg_ddlx pglinter pg_permissions safeupdate pg_dirtyread
      - pg_anon pgsmcrypto credcheck pg_vault pgsodium pg_session_jwt documentdb
      - pgsql-feat pgsql-type pgsql-util pgsql-func pgsql-fdw
    pg_users:
      - { name: dbuser_meta ,password: DBUser.Meta   ,pgbouncer: true ,roles: [dbrole_admin   ] ,comment: pigsty admin user }
      - { name: dbuser_view ,password: DBUser.Viewer ,pgbouncer: true ,roles: [dbrole_readonly] ,comment: read-only viewer  }
    pg_databases:
      - { name: meta, baseline: cmdb.sql ,comment: pigsty meta database ,schemas: [pigsty] ,extensions: [ postgis, timescaledb, vector ]}
    pg_libs: 'timescaledb, pg_stat_statements, auto_explain, pg_wait_sampling'
    pg_hba_rules:
      - { user: all ,db: all ,addr: intra ,auth: pwd ,title: 'everyone intranet access with password' ,order: 800 }
      - { user: all ,db: all ,addr: world ,auth: pwd ,title: 'everyone world access with password'    ,order: 900 }
    pg_crontab: [ '00 01 * * * /pg/bin/pg-backup full' ] # make a full backup every 1am
    patroni_mode: remove                # remove patroni after deployment
    pgbouncer_enabled: false            # disable pgbouncer pool
    pgbouncer_exporter_enabled: false   # disable pgbouncer_exporter on pgsql hosts?
    pgbackrest_exporter_enabled: false  # disable pgbackrest_exporter
    pg_default_services: []             # do not provision pg services
    #pg_reload: false                   # do not reload patroni/service

    #----------------------------------------------#
    # JUICE : https://pigsty.io/docs/juice
    #----------------------------------------------#
    juice_instances:  # dict of juicefs filesystems to deploy
      jfs:
        path  : /fs
        meta  : postgres://dbuser_meta:[email protected]:5432/meta
        data  : --storage postgres --bucket 10.10.10.10:5432/meta --access-key dbuser_meta --secret-key DBUser.Meta
        port  : 9567

    #----------------------------------------------#
    # VIBE : https://pigsty.io/docs/vibe
    #----------------------------------------------#
    # CHANAGE PASSWORD！
    code_password: Code.Server
    jupyter_password: Jupyter.Lab
    #claude_env:   # you can use other models here!
    #  ANTHROPIC_BASE_URL: https://open.bigmodel.cn/api/anthropic
    #  ANTHROPIC_API_URL: https://open.bigmodel.cn/api/anthropic
    #  ANTHROPIC_AUTH_TOKEN: your_api_service_token
    #  ANTHROPIC_MODEL: glm-4.7
    #  ANTHROPIC_SMALL_FAST_MODEL: glm-4.5-air

    #----------------------------------------------#
    # PASSWORD : https://pigsty.io/docs/setup/security/
    #----------------------------------------------#
    grafana_admin_password: pigsty
    grafana_view_password: DBUser.Viewer
    pg_admin_password: DBUser.DBA
    pg_monitor_password: DBUser.Monitor
    pg_replication_password: DBUser.Replicator
    patroni_password: Patroni.API
    haproxy_admin_password: pigsty
...

Explanation

The vibe template is an AI-era Web coding sandbox, enabling development, data analysis, AI app building all in browser.

Core Components:

Component	Description	Access Method
Code-Server	Web version of VS Code, full-featured code editor	`http://<ip>/code`
JupyterLab	Interactive data science notebook, Python/SQL	`http://<ip>/jupyter`
Claude Code	AI coding assistant CLI with OpenTelemetry observability	Terminal `claude` command
JuiceFS	PostgreSQL-based distributed filesystem	Mount point `/fs`
PostgreSQL 18	Feature-rich database with vector/timeseries/fulltext extensions	Port `5432`

Pre-installed Dev Tools:

AI Assistants: claude (Claude Code CLI), opencode (CLI AI coding tool)
Language Runtimes: golang, nodejs, uv (Python package manager)
Data Tools: postgrest (auto REST API), genai-toolbox
Utilities: restic, rclone (backup sync), asciinema (terminal recording)

PostgreSQL Extensions:

This template pre-installs rich PostgreSQL extensions covering AI/vector, timeseries, fulltext search, analytics:

# Vector & AI
pgvector, vchord, pgvectorscale, pg_search, pg_textsearch, vchord_bm25

# Timeseries & Geo
timescaledb, postgis, pg_cron

# Analytics & Lakehouse
pg_duckdb, pg_mooncake, pg_clickhouse, pg_parquet

# Security & Audit
pg_anon, pgsmcrypto, credcheck, pg_vault, pgsodium, pg_session_jwt

# Development
pg_tle, pljs, plprql, documentdb, pglinter

VIBE Module Components

VIBE module is new in v4.0.0, an AI coding sandbox module with three core components:

Code-Server: VS Code in browser

Full VS Code functionality, extension support
HTTPS access via Nginx reverse proxy
Supports Open VSX and Microsoft extension marketplaces
Related params: code_enabled, code_port, code_data, code_password, code_gallery

JupyterLab: Interactive computing environment

Python/SQL/Markdown notebook support
Pre-configured Python venv with data science libraries
HTTPS access via Nginx reverse proxy
Related params: jupyter_enabled, jupyter_port, jupyter_data, jupyter_password, jupyter_venv

Claude Code: AI coding assistant

Configure Claude Code CLI, skip initial onboarding
Built-in OpenTelemetry config, sends metrics/logs to Victoria stack
Provides claude-code dashboard for usage monitoring
Related params: claude_enabled, claude_env

JuiceFS Filesystem

This template uses JuiceFS for distributed filesystem capability, with a special feature: both metadata and data stored in PostgreSQL.

Architecture Features:

Metadata Engine: Uses PostgreSQL for filesystem metadata storage
Data Storage: Uses PostgreSQL Large Object for file data storage
Mount Point: Default mount at /fs (controlled by vibe_data param)
Monitoring Port: 9567 provides Prometheus metrics

Use Cases:

Persistent storage for code projects
Working directory for Jupyter Notebooks
Storage for AI models and datasets
File sharing across instances (when scaled to multiple nodes)

Config Example:

juice_instances:
  jfs:
    path  : /fs
    meta  : postgres://dbuser_meta:[email protected]:5432/meta
    data  : --storage postgres --bucket 10.10.10.10:5432/meta --access-key dbuser_meta --secret-key DBUser.Meta
    port  : 9567

Deployment Steps

# 1. Download Pigsty
curl https://repo.pigsty.io/get | bash

# 2. Use vibe config template
./configure -c vibe

# 3. Modify passwords (important!)
vi pigsty.yml
# Change code_password, jupyter_password, etc.

# 4. Deploy infrastructure and PostgreSQL
./deploy.yml

# 5. Deploy JuiceFS filesystem
./juice.yml

# 6. Deploy VIBE module (Code-Server, JupyterLab, Claude Code)
./vibe.yml

Access Methods

After deployment, access via browser:

# Code-Server (VS Code Web)
http://<ip>/code
# Password: Code.Server (please change)

# JupyterLab
http://<ip>/jupyter
# Password: Jupyter.Lab (please change)

# Claude Code Dashboard
http://<ip>:3000/d/claude-code
# Grafana default: admin / pigsty

# PostgreSQL
psql postgres://dbuser_meta:DBUser.Meta@<ip>:5432/meta

Use Cases

AI App Development: Build RAG, Agent, LLM applications
Data Science: Use JupyterLab for data analysis and visualization
Remote Development: Setup Web IDE environment on cloud servers
Teaching Demos: Provide consistent dev environment for students
Rapid Prototyping: Quickly validate ideas without local env setup
Claude Code Observability: Monitor AI coding assistant usage

Notes

Must change passwords: code_password and jupyter_password defaults are for testing only
Network security: This template opens world access (addr: world), production should configure firewall or VPN
Resource requirements: Recommend at least 2 cores 4GB memory, SSD disk
Simplified architecture: This template disables Patroni, PgBouncer etc HA components, suitable for single-node dev env
Claude API: Using Claude Code requires configuring API key in claude_env

Feedback

Was this page helpful?

Thanks for the feedback! Please let us know how we can improve.

Sorry to hear that. Please let us know how we can improve.

Last Modified 2026-01-25: v4.0 batch update (65761a0)