Quick Start

Prerequisites

Vibe Analyzer requires two external services:

OpenSearch — storage and search for indexed data
Ollama — running LLMs for code enrichment with descriptions and tags

Both services can be run locally via Docker.

Installation

Install via Cargo (recommended)

The package is available on crates.io:

cargo install vibe-analyzer

Build from Source

# 1. Clone the repository
git clone https://gitcode.com/keygenqt_vz/vibe-analyzer.git

# 2. Enter the directory
cd vibe-analyzer

# 3. Build
cargo build --release

Build Dependencies

Rust toolchain (cargo, rustc)
libssl-dev (for TLS)

Starting Services

The repository includes two ready-to-use docker-compose files:

docker/opensearch/docker-compose.yml — OpenSearch for indexing and search
docker/open-webui/docker-compose.yml — Open WebUI with Ollama for AI assistant connection

Docker (recommended)

OpenSearch:

cd docker/opensearch
docker-compose up -d

Open WebUI (optional, for AI assistant connection):

cd docker/open-webui
docker-compose up -d

Verification

# OpenSearch should respond
curl http://localhost:9200

# Ollama should be accessible
curl http://localhost:11434/api/tags

Configuration

The configuration file is located at ~/.vibe-analyzer/config.json5. It is created automatically with default settings the first time any CLI command is run.

Example Working Configuration

{
  // Configuration version (do not modify)
  "version": "0.0.1",

  // OpenSearch connection
  "opensearch": {
    "host": "http://192.168.1.10:9200"
  },

  // MCP server
  //
  // host — bind address (0.0.0.0 for all interfaces, 127.0.0.1 local only)
  // port — server port (default: 9020)
  // protocol — MCP protocol version (2024-11-05, 2025-03-26, 2025-06-18, or 'latest')
  "mcp": {
    "host": "0.0.0.0",
    "port": 9020,
    "protocol": "latest"
  },

  // Ollama LLM servers
  //
  // host — API endpoint
  // model — model name
  // max_chunk_chars — maximum characters per request
  // max_chunk_files — maximum files per request
  // timeout_secs — request timeout in seconds
  // temperature — generation temperature (0.0 – 1.0)
  // seed — seed for reproducibility
  // num_ctx — context window size
  // num_predict — maximum tokens in response
  // ast_imports — include imports in analysis
  // ast_variables — include variables in analysis
  // ast_functions — include functions in analysis
  // ast_enums — include enums in analysis
  // ast_interfaces — include interfaces in analysis
  "ollama": [
    {
      "host": "http://192.168.1.10:11434",
      "model": "qwen2.5-coder:3b-instruct",
      "max_chunk_chars": 4000,
      "max_chunk_files": 3,
      "timeout_secs": 60,
      "temperature": 0.1,
      "seed": 42,
      "num_ctx": 4096,
      "num_predict": 2048,
      "ast_imports": false,
      "ast_variables": false,
      "ast_functions": true,
      "ast_enums": true,
      "ast_interfaces": true
    },
    {
      "host": "http://localhost:11434",
      "model": "qwen2.5-coder:3b-instruct",
      "max_chunk_chars": 4000,
      "max_chunk_files": 3,
      "timeout_secs": 60,
      "temperature": 0.1,
      "seed": 42,
      "num_ctx": 4096,
      "num_predict": 2048,
      "ast_imports": false,
      "ast_variables": false,
      "ast_functions": true,
      "ast_enums": true,
      "ast_interfaces": true
    }
  ],

  // Knowledge sources for indexing
  "sources": ["/Users/keygenqt/Documents/Gitcode/Projects/vibe-analyzer"]
}

Checking Configuration

# View current settings
cat ~/.vibe-analyzer/config.json5

Adding a Knowledge Source

A source is anything you want to index: a code project, a documentation folder, or both.

# Add a project
vibe-analyzer source add /path/to/your/project

# Add a documentation directory
vibe-analyzer source add /path/to/docs

# List all added sources
vibe-analyzer source list

Scanning and Indexing

Vibe Analyzer provides three commands for different tasks:

Export AST to File

Code structure extraction only, without LLM. The result is saved in JSON/JSON5/TOON/XML:

# All sources
vibe-analyzer scan ast

# A specific source
vibe-analyzer scan ast --target /path/to/your/project

# With format specified
vibe-analyzer scan ast --target /path/to/your/project --format json5

Export AST with LLM Enrichment to File

AST parsing + enrichment via Ollama (descriptions, tags). The result is saved to a file:

vibe-analyzer scan analyze --target /path/to/your/project

Note: enrichment requires a running Ollama with the selected model. If multiple Ollama hosts are configured, files are distributed among them via competing consumers.

Indexing to OpenSearch

Full cycle — AST parsing, LLM enrichment, and writing to OpenSearch for search via MCP tools:

vibe-analyzer scan index --target /path/to/your/project

After indexing, data is ready for search through the MCP server.

Starting the MCP Server

vibe-analyzer serve start

The server starts on the address and port specified in the configuration (default http://0.0.0.0:9020).

Verifying Results

# Project statistics
vibe-analyzer stats info --target /path/to/your/project

# File tree
vibe-analyzer stats tree --target /path/to/your/project

# List all indexed projects
vibe-analyzer stats info

Connecting an AI Assistant

Open WebUI

Make sure Open WebUI is running (see the “Starting Services” section)
In Open WebUI settings, add an MCP server:
- URL: http://<host>:9020 (as specified in the configuration)
- Transport: Streamable HTTP
Once connected, the AI model will have 11 tools for searching code and documentation

Incremental Updates

Vibe Analyzer uses BLAKE3 hashes to track changes. When running scan index again, only modified files are processed:

# Reindexing — only changed files are affected
vibe-analyzer scan index --target /path/to/your/project

To force a full reindex, use the --force flag:

vibe-analyzer scan index --target /path/to/your/project --force

The same can be done via the admin_sync MCP tool without restarting the server.

Troubleshooting

OpenSearch Unreachable

# Check Docker container status
docker ps | grep opensearch

Ollama Not Responding

# Check if Ollama is running
curl http://localhost:11434/api/tags

Model Not Installed

# Download the model
ollama pull qwen2.5-coder:3b-instruct

What’s Next

Architecture — detailed breakdown of Vibe Analyzer’s internals
CLI Reference — complete command reference
Integrations — connecting AI assistants

Keyboard shortcuts

Vibe Analyzer