VibeCody Use Cases

VibeCody is an AI-powered coding assistant that runs anywhere you need it – on your laptop, in the cloud, or on a Raspberry Pi. Deploy it as a desktop app (VibeUI), a terminal companion (VibeCLI), or an always-on server (vibecli --serve --port 7878) that monitors your infrastructure, responds to webhooks, and runs scheduled tasks around the clock.

With 23 AI providers (from local Ollama to Claude, OpenAI, Gemini, and 20 more), 106+ REPL commands, 556+ skill files, and a full autonomous agent loop, VibeCody covers the entire software development lifecycle – from writing the first line of code to deploying, monitoring, and securing production systems. It also connects to Gmail/Outlook, Google/Outlook Calendar, Todoist, Notion, Jira, and Home Assistant for productivity workflows beyond coding.

See the Configuration Guide for setup and the Provider Guide for connecting your preferred AI backend.

Table of Contents

1. Developer Workflows
2. Always-On Automation
3. DevOps and Infrastructure
4. Productivity and Knowledge
5. Team Collaboration
6. Security and Compliance
7. Data and Analytics
8. IoT and Edge
9. Enterprise
10. Creative and Personal

1. Developer Workflows

AI-Powered Code Review

VibeCody’s code review engine runs 7 detectors (security/OWASP, complexity, style, documentation, tests, duplication, architecture) with 8-linter aggregation and quality gates. Get a prioritized list of issues with one command.

/aireview --path src/ --detectors all --quality-gate strict

Works on: all (cloud, desktop, Pi)

Automated Test Generation

Point the agent at a module and it generates unit tests, integration tests, and property-based tests with full coverage reporting. It reads your existing test patterns and matches the framework you already use.

/agent "Generate comprehensive tests for src/auth/login.rs targeting 90% coverage"

Works on: all

Intelligent Refactoring

The intent-aware refactoring engine analyzes your code, proposes structural improvements, and applies them in a single pass. It understands design patterns, SOLID principles, and language idioms.

/intent-refactor src/handlers/ --strategy extract-method --dry-run

Works on: all

Pull Request Automation

Generate PR descriptions with Mermaid diagrams, run pre-submit reviews, and auto-fix linter warnings before pushing. The agent reads the full diff, summarizes changes, and flags risks.

/agent "Create a PR for branch feature/auth-refactor with a summary, test plan, and risk assessment"

Works on: all

Interactive Debugging

Describe a bug in natural language. The agent reads stack traces, searches for related code, inserts diagnostic logging, runs the failing test, and proposes a fix – all in an autonomous loop with human approval gates.

/agent "The /api/users endpoint returns 500 when the email contains a plus sign. Find and fix the bug."

Works on: all

CI/CD Pipeline Management

Monitor GitHub Actions runs, view build logs, trigger workflows, and debug failing pipelines from VibeCLI or VibeUI. The GH Actions agent can diagnose flaky tests and suggest fixes.

/cicd status --repo myorg/myapp
/cicd logs --run 12345 --job build
/cicd trigger --workflow release.yml --ref main

Works on: all

Deployment Workflows

Deploy to cloud providers, Kubernetes clusters, or bare-metal servers with AI-assisted rollback detection. VibeCody generates deployment manifests, runs pre-flight checks, and monitors health after deploy.

/deploy --target k8s --namespace production --image myapp:v2.1.0 --strategy rolling

Works on: cloud, desktop

Git Workflow Automation

AI-assisted commit messages, interactive rebase planning, branch management, and conflict resolution. VibeCody reads the diff and writes conventional-commit-style messages.

/gitflow commit --conventional
/gitflow rebase-plan main..feature/auth

Works on: all

Autofix for Linter and Compiler Errors

Feed compiler errors or linter output to VibeCody and it applies fixes automatically, re-runs the check, and iterates until the build is clean.

/autofix --command "cargo clippy --workspace" --max-iterations 5

Works on: all

Code Replay and Explanation

Step through a git history range and have VibeCody explain each change, why it was made, and how it affects the rest of the system. Ideal for onboarding or post-incident review.

/code-replay --range HEAD~10..HEAD --annotate

Works on: all

Scaffold New Projects

Generate project scaffolding from templates – REST APIs, CLIs, microservices, React apps, and more – with CI config, Dockerfiles, and test harnesses included.

/scaffold --template rust-axum-api --name my-service --features auth,db,docker

Works on: all

2. Always-On Automation

Server Mode

Run VibeCody as a persistent HTTP daemon that accepts tasks via REST API. Integrate it into webhooks, chatbots, or custom tooling.

vibecli --serve --port 7878 --provider ollama
curl -X POST http://localhost:7878/task -d '{"prompt": "Review the latest commit"}'

Works on: cloud, desktop, Pi

Scheduled Agent Tasks

Define cron-style schedules for recurring tasks: nightly code quality scans, weekly dependency audits, or hourly health checks.

/schedule create --name "nightly-review" --cron "0 2 * * *" --command "/aireview --path src/"
/schedule list

Works on: cloud, Pi

Auto-Research

VibeCody autonomously researches a topic, crawls documentation, and produces a structured report. Leave it running overnight and read the results in the morning.

/auto-research "Compare WebSocket libraries for Rust: tokio-tungstenite vs axum vs actix-web"

Works on: all

Health Monitoring

Continuously monitor application health endpoints, log response times, and alert when services degrade. The AI can diagnose issues and suggest remediations.

/health-monitor --endpoints config/endpoints.json --interval 60s --alert slack

Works on: cloud, Pi

Webhook-Driven Pipelines

Configure VibeCody’s server mode to respond to GitHub webhooks: auto-review PRs on open, run security scans on push, or generate release notes on tag.

# In config.toml
[hooks.on_pr_opened]
command = "/aireview --pr ${PR_NUMBER}"

[hooks.on_push]
command = "/redteam --changed-files-only"

Works on: cloud, Pi

Dependency Update Monitoring

Schedule periodic dependency audits that check for CVEs, outdated packages, and license compliance. VibeCody creates PRs for safe updates.

/schedule create --name "dep-audit" --cron "0 8 * * 1" --command "/agent 'Audit dependencies, update safe patches, create a PR'"

Works on: cloud, Pi

Log Analysis

Point VibeCody at log files or streams and have it identify anomalies, correlate errors across services, and surface root causes.

/agent "Analyze /var/log/app/*.log from the last 24 hours. Find error patterns and suggest fixes."

Works on: cloud, desktop, Pi

Automated Documentation Sync

Schedule doc generation from code comments, API schemas, and architecture decision records. Keep your docs site in sync with your codebase automatically.

/schedule create --name "doc-sync" --cron "0 6 * * *" --command "/agent 'Regenerate API docs from OpenAPI spec and update docs/ folder'"

Works on: cloud, Pi

Environment Drift Detection

Compare staging and production configs on a schedule and alert when drift is detected. The agent generates a remediation plan.

/schedule create --name "drift-check" --cron "0 */4 * * *" --command "/env-diff staging production --alert"

Works on: cloud, Pi

3. DevOps and Infrastructure

Docker Container Management

Build, run, inspect, and debug Docker containers directly from VibeCLI. VibeCody generates Dockerfiles, docker-compose configs, and multi-stage builds optimized for your stack.

/docker build --context . --target production --optimize
/docker compose up --detach
/docker logs my-service --tail 100 --analyze

Works on: cloud, desktop

Kubernetes Operations

Deploy, scale, and troubleshoot Kubernetes workloads. VibeCody reads cluster state, generates manifests, and applies rolling updates.

/k8s status --namespace production
/k8s deploy --manifest k8s/deployment.yaml --dry-run
/k8s debug pod/my-app-7d4f8b --logs --events

Works on: cloud, desktop

Cloud Provider Scanning

Scan your AWS, GCP, or Azure accounts for misconfigured resources, cost optimization opportunities, and security issues.

/cloud scan --provider aws --checks security,cost,compliance
/cloud report --format markdown

Works on: cloud, desktop

Infrastructure as Code Generation

Describe your desired infrastructure in natural language and VibeCody generates Terraform, Pulumi, CloudFormation, or CDK code.

/agent "Generate Terraform for a 3-tier AWS architecture: ALB, ECS Fargate, Aurora PostgreSQL, with VPC and security groups"

Works on: all

Cost Estimation

Estimate cloud costs before deploying. VibeCody analyzes your IaC files and produces a monthly cost breakdown per resource.

/cloud cost-estimate --terraform infra/ --region us-east-1

Works on: all

Network Diagnostics

Run DNS lookups, traceroutes, port scans, and TLS certificate checks from VibeCody. The AI interprets results and suggests fixes.

/network diagnose --target api.example.com --checks dns,tls,latency,ports

Works on: cloud, desktop, Pi

SSH Session Management

Connect to remote servers, run commands, and transfer files through VibeCody’s SSH integration. The agent can troubleshoot remote issues interactively.

/ssh connect production-web-01 --command "systemctl status nginx"
/ssh exec-all --group web-servers --command "df -h"

Works on: cloud, desktop

CI/CD Pipeline Generation

Generate GitHub Actions, GitLab CI, or Jenkins pipelines from your project structure. VibeCody detects the language, test framework, and deployment target automatically.

/agent "Generate a GitHub Actions workflow for this Rust project: lint, test, build Docker image, deploy to ECS"

Works on: all

Container Sandbox Execution

Run agent tasks in isolated Docker or Podman containers for safety. The sandbox supports 16 async operations with resource limits and network policies.

vibecli --sandbox docker --agent "Refactor the payment module"

Works on: cloud, desktop

Load Testing

Generate and run load test scenarios against your APIs. VibeCody creates test scripts, executes them, and analyzes the results.

/loadtest --target http://localhost:8080/api --rps 500 --duration 60s --analyze

Works on: cloud, desktop

4. Productivity and Knowledge

Context Bundles

Save and restore conversation contexts as reusable bundles. Share a bundle with teammates so they can pick up where you left off.

/context save auth-refactor "Context for the authentication rewrite project"
/context load auth-refactor
/context list

Works on: all

Open Memory

VibeCody maintains an auto-updated memory of your project – architecture decisions, coding patterns, team conventions – that persists across sessions and improves suggestions over time.

/memory show
/memory add "We use snake_case for database columns and camelCase for API responses"
/memory search "authentication flow"

Works on: all

Session Management

Create, resume, and branch sessions backed by SQLite. Each session preserves full conversation history, tool outputs, and agent state.

/session new --name "bug-1234"
/session resume bug-1234
/session list --last 10
/session branch bug-1234 --name "bug-1234-alt-approach"

Works on: all

AI-Assisted Documentation

Generate README files, API docs, architecture decision records, and inline documentation from your codebase.

/agent "Generate API documentation for all endpoints in src/routes/ using OpenAPI 3.1 format"

Works on: all

Codebase Q&A

Ask questions about your codebase in natural language. VibeCody searches files, reads relevant code, and answers with citations.

vibecli -c "How does the authentication middleware work? Which files are involved?"

Works on: all

Bookmark and Snippet Management

Save code snippets, file locations, and search results as bookmarks for quick reference later.

/bookmark add src/auth/middleware.rs:42 "JWT validation logic"
/bookmark list --tag auth
/snippet save "retry-with-backoff" --file src/utils/retry.rs --lines 10-35

Works on: all

Architecture Decision Records

Create and manage ADRs with lifecycle tracking, markdown generation, and governance rules powered by the TOGAF/Zachman engine.

/archspec adr create "Use PostgreSQL for user data" --status accepted --context "Evaluated Mongo, DynamoDB, Postgres"
/archspec adr list --status accepted

Works on: all

One-Shot Queries

Get quick answers without entering the REPL. Pipe output from other tools for instant analysis.

vibecli -c "Explain what this regex does: ^(?:[a-z0-9!#$%&'*+/=?^_\`{|}~-]+\.)*"
cat error.log | vibecli -c "What caused this crash?"

Works on: all

Profile Switching

Maintain separate profiles for different projects, teams, or contexts. Each profile has its own provider, model, system prompt, and rules.

vibecli --profile work
vibecli --profile personal
/profile list
/profile create oss --provider ollama --model codellama:34b

Works on: all

Email Triage and Management

Connect Gmail or Outlook and manage email from the terminal. The AI triage mode classifies messages as urgent, action-needed, or FYI and suggests responses.

# Morning briefing
/email unread           # 14 unread messages
/email triage           # AI classifies + archives FYI messages

# Send a reply
/email read <id>
/email send alice@co.com "Re: deploy plan" "Looks good — approved"

Configure in ~/.vibecli/config.toml under [email] or set GMAIL_ACCESS_TOKEN / OUTLOOK_ACCESS_TOKEN.

Works on: all

Calendar and Scheduling

View and manage Google Calendar or Outlook Calendar events. Find free slots before proposing meetings.

/cal today              # Today's events with times
/cal free tomorrow      # Open slots tomorrow
/cal create "Sprint review" "friday 2pm" "friday 3pm"
/cal next               # Next upcoming event

Works on: all

Task Management with Todoist

Add, complete, and triage tasks without leaving the terminal. Integrates with Jira for cross-system tracking.

/todo today             # Tasks due today + overdue
/todo add "Fix login bug" due:today p1
/todo close 1234567     # Mark done
/todo project "Work"    # View project tasks

Works on: all

Notion Knowledge Base

Search, read, and append to Notion pages without opening a browser. Useful for logging meeting notes and cross-referencing docs during development.

/notion search "API design decisions"
/notion get abc123def456    # Read page content
/notion append abc123def456 "2026-04-04: switched to REST from GraphQL"

Works on: all

Jira Issue Tracking

Create tickets, update statuses, and add comments from the terminal. Works with Jira Cloud and Server.

/jira mine              # My open issues
/jira create PROJ "Null pointer in UserService" "Stack: ..."
/jira transition PROJ-234 "In Review"
/jira comment PROJ-234 "Root cause found — missing null check on line 47"

Works on: all

Smart Home Control via Home Assistant

Control lights, thermostats, scenes, and automations from the terminal. Run focus or wind-down routines as part of your work workflow.

/ha scene focus         # Dim lights, set temperature for deep work
/ha status              # Full home state summary
/ha off all lights      # Lights out
/ha climate thermostat.main 72

Requires Home Assistant with a long-lived access token. Works locally or via Tailscale/Nabu Casa for remote access.

Works on: all

Morning Briefing Sequence

Run a complete morning briefing in four commands — no browser tabs needed.

/email unread     # Overnight emails
/cal today        # Today's meetings
/todo today       # Tasks due today
/jira mine        # Open Jira issues

Works on: all

5. Team Collaboration

Agent Teams

Spin up multi-agent teams with specialized roles (Architect, Coder, Reviewer, Tester, SecurityAuditor) that collaborate on tasks through an inter-agent messaging bus.

/team create feature-auth --roles architect,coder,reviewer,tester
/team run feature-auth "Design and implement OAuth2 PKCE flow for the mobile app"

Works on: cloud, desktop

CRDT Collaborative Editing

Multiple users can edit the same file simultaneously with conflict-free resolution using CRDTs. Changes are synchronized in real time over Tailscale or local networks.

/collab start --file src/main.rs --peers alice@tailscale,bob@tailscale
/collab status

Works on: cloud, desktop

Gateway Messaging (18 Platforms)

Connect VibeCody to Slack, Discord, Teams, Telegram, WhatsApp, and 13 other platforms. Team members can interact with VibeCody from their preferred messaging tool.

vibecli --serve --port 7878 --gateway slack,discord,teams
/gateway status
/gateway send slack "#engineering" "Deploy v2.1 completed successfully"

Works on: cloud, Pi

Multi-Agent Orchestration

Coordinate multiple agents working on different parts of a codebase. The orchestrator handles task decomposition, dependency resolution, and result synthesis.

/orchestrate "Refactor the payment system" --agents 4 --strategy divide-and-conquer

Works on: cloud, desktop

Shared Agent Memory

Agent teams share a memory store so discoveries by one agent (e.g., a security flaw found by the auditor) are immediately available to all other agents on the team.

/team memory show feature-auth
/team memory add feature-auth "The existing session table uses UUID v4 primary keys"

Works on: cloud, desktop

Voice Pairing via Tailscale

Pair-program with VibeCody using voice commands over a Tailscale connection. Speak your intent, and VibeCody writes the code.

/voice start --tailscale --device my-laptop
/voice language en-US
# "Add error handling to the upload function"

Works on: desktop, Pi

Code Review Protocol

Define team review standards as machine-readable protocols. VibeCody enforces them automatically on every PR, ensuring consistent quality.

/review-protocol load team-standards.yaml
/review-protocol run --pr 456 --strict

Works on: all

Linear and Jira Integration

Create, update, and track issues directly from VibeCLI. The agent can read issue descriptions and automatically start working on them.

/linear list --status "In Progress" --assignee me
/agent "Pick up LINEAR-1234 and implement the feature described in the ticket"

Works on: all

6. Security and Compliance

Red Team Security Testing

Run automated adversarial security scans against your codebase. VibeCody probes for OWASP Top 10 vulnerabilities, path traversal, SQL injection, XSS, and insecure deserialization.

/redteam scan --path src/ --checks owasp-top-10 --severity high,critical
/redteam report --format sarif

Works on: all

OWASP Vulnerability Scanning

Targeted scanning for OWASP categories with remediation suggestions that the agent can apply automatically.

/redteam owasp --category injection,xss,auth-failure --autofix

Works on: all

Policy Engine

Define RBAC/ABAC policies in YAML (Cerbos-style) with 14 condition operators, derived roles, conflict detection, and coverage analysis. Enforce who can do what across your system.

/policy load policies/
/policy test --suite policy-tests.yaml
/policy audit --last 24h

Works on: all

Architecture Governance

Enforce architectural boundaries with a TOGAF/Zachman governance engine. Detect violations like circular dependencies, layer breaches, and unauthorized direct database access from controllers.

/archspec governance check --rules arch-rules.yaml
/archspec c4 generate --level component --output docs/architecture.md

Works on: all

Dependency Vulnerability Audit

Scan dependencies across Rust (cargo-audit), Node (npm audit), Python (pip-audit), and more. VibeCody correlates CVEs with your actual usage to eliminate false positives.

/agent "Audit all dependencies for known vulnerabilities. For each CVE, check if our code actually calls the affected function."

Works on: all

Secret Detection

Scan your codebase and git history for leaked secrets: API keys, passwords, tokens, and private keys. VibeCody identifies the commit, author, and suggests rotation steps.

/redteam secrets --scan-history --depth 100

Works on: all

Compliance Reporting

Generate compliance reports for SOC2, HIPAA, PCI-DSS, and GDPR. VibeCody maps your codebase controls to compliance frameworks and identifies gaps.

/policy compliance --framework soc2 --output compliance-report.md

Works on: all

Blue/Purple Team Exercises

Run coordinated attack-and-defend exercises where one agent team attacks and another defends. Results are scored and reported.

/purpleteam run --attack-surface api --duration 30m --report

Works on: cloud, desktop

Audit Trail

Every agent action is logged to JSONL trace files with context sidecars. Full audit trail for compliance and post-incident review.

/trace list --last 7d
/trace show 2026-04-01T14-30-00 --include-context
/trace export --format csv --output audit.csv

Works on: all

7. Data and Analytics

Database Management

Connect to PostgreSQL, MySQL, SQLite, MongoDB, Redis, and DuckDB. Run queries, inspect schemas, generate migrations, and optimize slow queries with AI assistance.

/db connect postgres://user:pass@localhost:5432/mydb
/db schema --table users
/db query "SELECT * FROM orders WHERE status = 'pending' LIMIT 10"
/db optimize --slow-query-log

Works on: all

Cost Observatory

Track AI token usage, API costs, and cloud spend across all 23 providers. Set budgets and alerts to prevent surprise bills.

/cost show --period this-month --by-provider
/cost budget set --monthly 50.00 --alert-at 80%
/cost export --format csv

Works on: all

Embeddings and Semantic Search

Generate embeddings for your codebase and use semantic search to find code by meaning rather than keywords. Powered by optimized O(n) updates and fused cosine similarity.

/embeddings index --path src/
/embeddings search "function that handles user authentication with JWT"

Works on: all

Metrics Collection and Visualization

Collect application metrics, analyze trends, and generate dashboards. VibeCody integrates with Prometheus, Grafana, and custom metric sources.

/metrics collect --source prometheus --query 'http_requests_total{status="500"}'
/metrics analyze --period 7d --anomaly-detection

Works on: cloud, desktop

Database Migration Generation

Describe schema changes in natural language and VibeCody generates migration files compatible with your ORM (Diesel, SQLx, Prisma, Alembic, etc.).

/agent "Add a 'teams' table with id, name, created_at. Add team_id foreign key to users. Generate a Diesel migration."

Works on: all

GraphQL Schema Management

Generate, validate, and evolve GraphQL schemas. VibeCody creates resolvers, types, and mutations from your database schema.

/graphql generate --from-db postgres://localhost/mydb --output schema.graphql
/graphql validate schema.graphql

Works on: all

API Performance Analysis

Analyze API response times, error rates, and throughput from logs or monitoring data. VibeCody identifies bottlenecks and suggests optimizations.

/agent "Analyze the nginx access logs for the last 24 hours. Which endpoints are slowest? What is the p99 latency?"

Works on: all

Data Pipeline Debugging

Trace data through ETL pipelines, identify where transformations fail, and suggest fixes. VibeCody reads pipeline configs and log outputs.

/agent "The daily user_events ETL pipeline failed at the transform step. Diagnose from the Airflow logs and fix the SQL."

Works on: cloud, desktop

8. IoT and Edge

Raspberry Pi Deployment

VibeCody compiles to a single ARM binary that runs on Raspberry Pi. Use it as a local AI assistant, home automation controller, or edge compute node.

# Install on Raspberry Pi
curl -fsSL https://vibecody.dev/install.sh | sh
vibecli --provider ollama --model tinyllama:1.1b

Works on: Pi

Smart Home Automation via MCP

Connect to smart home devices through MCP (Model Context Protocol) servers. Control lights, thermostats, sensors, and appliances with natural language.

# Connect to Home Assistant MCP server
/mcp connect homeassistant --url http://homeassistant.local:8123/mcp

# Natural language control
/agent "Turn off all lights except the living room. Set thermostat to 68F."

Works on: Pi, desktop

Edge Sensor Monitoring

Read sensor data from GPIO, I2C, or MQTT sources. VibeCody analyzes patterns and triggers alerts when thresholds are exceeded.

/agent "Monitor the MQTT topic sensors/temperature. Alert me if it exceeds 85F for more than 5 minutes."

Works on: Pi

Local-Only AI Processing

Run entirely offline with Ollama and a local model. No data leaves your device. Ideal for air-gapped environments and privacy-sensitive workloads.

vibecli --provider ollama --model codellama:13b --offline

Works on: Pi, desktop

Edge Deployment Orchestration

Deploy and update applications across a fleet of edge devices. VibeCody generates OTA update manifests and manages rollback strategies.

/agent "Generate a Balena fleet deployment config for our sensor-reader app targeting RPi4 devices"

Works on: cloud, Pi

Voice-Controlled Development

Use voice commands to write code, run tests, and manage deployments hands-free. Useful for hardware prototyping when your hands are occupied.

/voice start --provider whisper --device default
# "Run the test suite and tell me if anything failed"

Works on: Pi, desktop

MCP Server Integration

Connect to any MCP-compatible tool server to extend VibeCody’s capabilities. Browse the MCP directory and lazy-load servers on demand.

/mcp directory search "database"
/mcp connect sqlite --lazy
/mcp list

Works on: all

Kiosk and Display Mode

Run VibeUI in kiosk mode on a dedicated display for dashboards, monitoring, or team status boards.

vibeui --kiosk --panel cost-observatory --refresh 60s

Works on: Pi, desktop

9. Enterprise

Multi-Provider Failover

Configure automatic failover between providers. If Claude is down, VibeCody seamlessly switches to OpenAI, then Gemini, with no interruption.

# In config.toml
[failover]
chain = ["claude", "openai", "gemini", "ollama"]
timeout_ms = 5000
retry_count = 2

Works on: all

Bring Your Own Key (BYOK)

Every team member uses their own API keys. No shared keys, no central billing proxy. Keys can be stored in vaults, env vars, or helper scripts.

# Key from a vault helper
[claude]
api_key_helper = "vault read -field=key secret/anthropic"

Works on: all

Role-Based Access Control

Define granular permissions for who can run which commands, access which tools, and modify which files using the policy engine.

/policy create --role junior-dev --allow "read,test,lint" --deny "deploy,delete,db-write"
/policy assign --user alice --role junior-dev

Works on: all

Audit and Compliance Logging

Every agent interaction produces JSONL traces with full context. Export to SIEM systems for enterprise compliance requirements.

/trace export --format siem --destination splunk://logs.corp.com:8088

Works on: all

Gateway for 18 Messaging Platforms

Deploy VibeCody as a shared team resource accessible from Slack, Teams, Discord, Telegram, and 14 more platforms. Centralized AI assistant for the entire organization.

vibecli --serve --port 7878 --gateway all --auth-mode oauth2

Works on: cloud

Profile-Based Multi-Tenancy

Separate configurations for different teams, projects, or environments. Each profile has isolated sessions, memory, and provider settings.

/profile create team-backend --provider claude --model claude-opus-4-6 --rules strict
/profile create team-frontend --provider openai --model gpt-4o --rules relaxed

Works on: all

Self-Hosted Deployment

Run VibeCody entirely on your own infrastructure. No SaaS dependency, no data exfiltration risk. Ship the single binary or use the Docker image.

docker run -d -p 7878:7878 -v /data:/data vibecody/vibecli:latest --serve

Works on: cloud

Custom Rules and Guardrails

Define project-specific rules that the agent must follow: coding standards, forbidden patterns, required review steps, and more.

# .vibecli/rules.toml
[[rules]]
name = "no-unwrap-in-prod"
pattern = "unwrap()"
action = "block"
message = "Use proper error handling instead of unwrap() in production code"

Works on: all

SSO and Token Management

Integrate with enterprise SSO providers for authentication. API token rotation, expiry tracking, and BYOK management are built in.

/tokens list --provider all
/tokens rotate --provider openai --confirm
/tokens usage --period this-month

Works on: all

10. Creative and Personal

Model Arena

Compare responses from multiple AI providers side-by-side. Run the same prompt through Claude, GPT-4o, Gemini, and local models to find the best answer.

/arena --providers claude,openai,gemini,ollama --prompt "Write a retry function with exponential backoff in Rust"

Works on: all

Inline Chat Editing

Edit code interactively in your terminal. Select a range of lines and describe the change you want. VibeCody applies the edit in place.

/inline src/main.rs:42-58 "Add proper error handling with anyhow context"

Works on: all

One-Shot Chat

Quick answers without entering the REPL. Great for shell aliases, editor integrations, and scripts.

vibecli -c "What is the time complexity of Rust's BTreeMap::insert?"
echo "SELECT * FORM users" | vibecli -c "Fix the SQL syntax error"

Works on: all

Voice Pair Programming

Speak naturally to VibeCody while coding. It listens, understands context from your open files, and writes code or runs commands based on your voice input.

/voice start --continuous
# "Add a health check endpoint that returns the git commit hash and uptime"

Works on: desktop, Pi

Personal Knowledge Base

Use VibeCody as a persistent personal knowledge assistant. Store notes, code snippets, and research across sessions with full semantic search.

/memory add "Rust lifetimes: &'a means the reference lives at least as long as 'a"
/memory search "lifetimes"
/snippet save "tokio-spawn-pattern" --from-clipboard

Works on: all

Learning and Exploration

Ask VibeCody to explain unfamiliar codebases, libraries, or concepts. It reads source code, documentation, and examples to teach you.

/agent "I am new to this project. Walk me through the architecture, entry points, and key data flows."

Works on: all

Blog and Technical Writing

Generate technical blog posts, tutorials, and documentation from your code and commit history.

/agent "Write a blog post explaining how our new authentication system works, with code examples from the actual implementation"

Works on: all

Git Bisect Assistance

VibeCody automates git bisect by running test commands and binary-searching for the commit that introduced a bug.

/gitflow bisect --good v2.0.0 --bad HEAD --test "cargo test auth_tests"

Works on: all

HTTP Playground

Test APIs interactively with an AI-assisted HTTP client. VibeCody reads your API schema and generates request examples.

/http GET http://localhost:8080/api/users --auth bearer
/http POST http://localhost:8080/api/users --body '{"name":"Alice"}' --analyze-response

Works on: all

Speculative Execution

VibeCody predicts what you will ask next and pre-computes likely responses using TurboQuant KV-cache. Results appear instantly when you confirm.

/speculative on
# VibeCody pre-generates the next likely edit after each interaction

Works on: desktop

Comparison: VibeCody vs myclaw.ai / OpenClaw

Getting Started

Ready to try these use cases? Start here:

1. Easy Setup: Run vibecli --setup or see the Easy Setup Guide
2. Install VibeCody: curl -fsSL https://raw.githubusercontent.com/TuringWorks/vibecody/main/install.sh | sh
3. Configure a provider: See the Provider Guide
4. Run your first agent task: vibecli --agent "Hello, review my project structure"
5. Explore REPL commands: Type /help in the REPL
6. Deploy anywhere: See the Deployment Guides for 12 platforms (AWS, GCP, Azure, Pi, etc.)