RLAMA is a powerful AI-driven question-answering tool for your documents, seamlessly integrating with your local Ollama models. It enables you to create, manage, and interact with Retrieval-Augmented Generation (RAG) systems tailored to your documentation needs.
- Vision & Roadmap
- Installation
- Available Commands
- rag - Create a RAG system
- crawl-rag - Create a RAG system from a website
- wizard - Create a RAG system with interactive setup
- watch - Set up directory watching for a RAG system
- watch-off - Disable directory watching for a RAG system
- check-watched - Check a RAG's watched directory for new files
- run - Use a RAG system
- api - Start API server
- list - List RAG systems
- delete - Delete a RAG system
- list-docs - List documents in a RAG
- list-chunks - Inspect document chunks
- view-chunk - View chunk details
- add-docs - Add documents to RAG
- crawl-add-docs - Add website content to RAG
- update-model - Change LLM model
- update - Update RLAMA
- version - Display version
- Uninstallation
- Supported Document Formats
- Troubleshooting
RLAMA aims to become the definitive tool for creating local RAG systems that work seamlessly for everyone—from individual developers to large enterprises. Here's our strategic roadmap:
- ✅ Basic RAG System Creation: CLI tool for creating and managing RAG systems
- ✅ Document Processing: Support for multiple document formats (.txt, .md, .pdf, etc.)
- ✅ Document Chunking: Basic text splitting with configurable size and overlap
- ✅ Vector Storage: Local storage of document embeddings
- ✅ Context Retrieval: Basic semantic search with configurable context size
- ✅ Ollama Integration: Seamless connection to Ollama models
- ✅ Cross-Platform Support: Works on Linux, macOS, and Windows
- ✅ Easy Installation: One-line installation script
- ✅ API Server: HTTP endpoints for integrating RAG capabilities in other applications
- ✅ Web Crawling: Create RAGs directly from websites
- ✅ Guided RAG Setup Wizard: Interactive interface for easy RAG creation
- Prompt Compression: Smart context summarization for limited context windows
- Adaptive Chunking: Dynamic content segmentation based on semantic boundaries
- Minimal Context Retrieval: Intelligent filtering to eliminate redundant content
- Parameter Optimization: Fine-tuned settings for different model sizes
- Multi-Model Embedding Support: Integration with various embedding models
- Hybrid Retrieval Techniques: Combining sparse and dense retrievers for better accuracy
- Embedding Evaluation Tools: Built-in metrics to measure retrieval quality
- Automated Embedding Cache: Smart caching to reduce computation for similar queries
- Lightweight Web Interface: Simple browser-based UI for the existing CLI backend
- Knowledge Graph Visualization: Interactive exploration of document connections
- Domain-Specific Templates: Pre-configured settings for different domains
- Multi-User Access Control: Role-based permissions for team environments
- Integration with Enterprise Systems: Connectors for SharePoint, Confluence, Google Workspace
- Knowledge Quality Monitoring: Detection of outdated or contradictory information
- System Integration API: Webhooks and APIs for embedding RLAMA in existing workflows
- AI Agent Creation Framework: Simplified system for building custom AI agents with RAG capabilities
- Multi-Step Retrieval: Using the LLM to refine search queries for complex questions
- Cross-Modal Retrieval: Support for image content understanding and retrieval
- Feedback-Based Optimization: Learning from user interactions to improve retrieval
- Knowledge Graphs & Symbolic Reasoning: Combining vector search with structured knowledge
RLAMA's core philosophy remains unchanged: to provide a simple, powerful, local RAG solution that respects privacy, minimizes resource requirements, and works seamlessly across platforms.
- Ollama installed and running
curl -fsSL https://raw.githubusercontent.com/dontizi/rlama/main/install.sh | shRLAMA is built with:
- Core Language: Go (chosen for performance, cross-platform compatibility, and single binary distribution)
- CLI Framework: Cobra (for command-line interface structure)
- LLM Integration: Ollama API (for embeddings and completions)
- Storage: Local filesystem-based storage (JSON files for simplicity and portability)
- Vector Search: Custom implementation of cosine similarity for embedding retrieval
RLAMA follows a clean architecture pattern with clear separation of concerns:
rlama/
├── cmd/ # CLI commands (using Cobra)
│ ├── root.go # Base command
│ ├── rag.go # Create RAG systems
│ ├── run.go # Query RAG systems
│ └── ...
├── internal/
│ ├── client/ # External API clients
│ │ └── ollama_client.go # Ollama API integration
│ ├── domain/ # Core domain models
│ │ ├── rag.go # RAG system entity
│ │ └── document.go # Document entity
│ ├── repository/ # Data persistence
│ │ └── rag_repository.go # Handles saving/loading RAGs
│ └── service/ # Business logic
│ ├── rag_service.go # RAG operations
│ ├── document_loader.go # Document processing
│ └── embedding_service.go # Vector embeddings
└── pkg/ # Shared utilities
└── vector/ # Vector operations
- Document Processing: Documents are loaded from the file system, parsed based on their type, and converted to plain text.
- Embedding Generation: Document text is sent to Ollama to generate vector embeddings.
- Storage: The RAG system (documents + embeddings) is stored in the user's home directory (~/.rlama).
- Query Process: When a user asks a question, it's converted to an embedding, compared against stored document embeddings, and relevant content is retrieved.
- Response Generation: Retrieved content and the question are sent to Ollama to generate a contextually-informed response.
┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ Documents │────>│ Document │────>│ Embedding │
│ (Input) │ │ Processing │ │ Generation │
└─────────────┘ └─────────────┘ └─────────────┘
│
▼
┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ Query │────>│ Vector │<────│ Vector Store│
│ Response │ │ Search │ │ (RAG System)│
└─────────────┘ └─────────────┘ └─────────────┘
▲ │
│ ▼
┌─────────────┐ ┌─────────────┐
│ Ollama │<────│ Context │
│ LLM │ │ Building │
└─────────────┘ └─────────────┘
RLAMA is designed to be lightweight and portable, focusing on providing RAG capabilities with minimal dependencies. The entire system runs locally, with the only external dependency being Ollama for LLM capabilities.
You can get help on all commands by using:
rlama --helpThese flags can be used with any command:
--host string Ollama host (default: localhost)
--port string Ollama port (default: 11434)Creates a new RAG system by indexing all documents in the specified folder.
rlama rag [model] [rag-name] [folder-path]Parameters:
model: Name of the Ollama model to use (e.g., llama3, mistral, gemma).rag-name: Unique name to identify your RAG system.folder-path: Path to the folder containing your documents.
Example:
rlama rag llama3 documentation ./docsCreates a new RAG system by crawling a website and indexing its content.
rlama crawl-rag [model] [rag-name] [website-url]Parameters:
model: Name of the Ollama model to use (e.g., llama3, mistral, gemma).rag-name: Unique name to identify your RAG system.website-url: URL of the website to crawl and index.
Options:
--max-depth: Maximum crawl depth (default: 2)--concurrency: Number of concurrent crawlers (default: 5)--exclude-path: Paths to exclude from crawling (comma-separated)--chunk-size: Character count per chunk (default: 1000)--chunk-overlap: Overlap between chunks in characters (default: 200)
Example:
# Create a new RAG from a documentation website
rlama crawl-rag llama3 docs-rag https://docs.example.com
# Customize crawling behavior
rlama crawl-rag llama3 blog-rag https://blog.example.com --max-depth=3 --exclude-path=/archive,/tagsProvides an interactive step-by-step wizard for creating a new RAG system.
rlama wizardThe wizard guides you through:
- Naming your RAG
- Choosing an Ollama model
- Selecting document sources (local folder or website)
- Configuring chunking parameters
- Setting up file filtering
Example:
rlama wizard
# Follow the prompts to create your customized RAGConfigure a RAG system to automatically watch a directory for new files and add them to the RAG.
rlama watch [rag-name] [directory-path] [interval]Parameters:
rag-name: Name of the RAG system to watch.directory-path: Path to the directory to watch for new files.interval: Time in minutes to check for new files (use 0 to check only when the RAG is used).
Example:
# Set up directory watching to check every 60 minutes
rlama watch my-docs ./watched-folder 60
# Set up directory watching to only check when the RAG is used
rlama watch my-docs ./watched-folder 0
# Customize what files to watch
rlama watch my-docs ./watched-folder 30 --exclude-dir=node_modules,tmp --process-ext=.md,.txtDisable automatic directory watching for a RAG system.
rlama watch-off [rag-name]Parameters:
rag-name: Name of the RAG system to disable watching.
Example:
rlama watch-off my-docsManually check a RAG's watched directory for new files and add them to the RAG.
rlama check-watched [rag-name]Parameters:
rag-name: Name of the RAG system to check.
Example:
rlama check-watched my-docsStarts an interactive session to interact with an existing RAG system.
rlama run [rag-name]Parameters:
rag-name: Name of the RAG system to use.--context-size: (Optional) Number of context chunks to retrieve (default: 20)
Example:
rlama run documentation
> How do I install the project?
> What are the main features?
> exitContext Size Tips:
- Smaller values (5-15) for faster responses with key information
- Medium values (20-40) for balanced performance
- Larger values (50+) for complex questions needing broad context
- Consider your model's context window limits
rlama run documentation --context-size=50 # Use 50 context chunksStarts an HTTP API server that exposes RLAMA's functionality through RESTful endpoints.
rlama api [--port PORT]Parameters:
--port: (Optional) Port number to run the API server on (default: 11249)
Example:
rlama api --port 8080Available Endpoints:
-
Query a RAG system -
POST /ragcurl -X POST http://localhost:11249/rag \ -H "Content-Type: application/json" \ -d '{ "rag_name": "documentation", "prompt": "How do I install the project?", "context_size": 20 }'
Request fields:
rag_name(required): Name of the RAG system to queryprompt(required): Question or prompt to send to the RAGcontext_size(optional): Number of chunks to include in contextmodel(optional): Override the model used by the RAG
-
Check server health -
GET /healthcurl http://localhost:11249/health
Integration Example:
// Node.js example
const response = await fetch('http://localhost:11249/rag', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({
rag_name: 'my-docs',
prompt: 'Summarize the key features'
})
});
const data = await response.json();
console.log(data.response);Displays a list of all available RAG systems.
rlama listPermanently deletes a RAG system and all its indexed documents.
rlama delete [rag-name] [--force/-f]Parameters:
rag-name: Name of the RAG system to delete.--forceor-f: (Optional) Delete without asking for confirmation.
Example:
rlama delete old-projectOr to delete without confirmation:
rlama delete old-project --forceDisplays all documents in a RAG system with metadata.
rlama list-docs [rag-name]Parameters:
rag-name: Name of the RAG system
Example:
rlama list-docs documentationList and filter document chunks in a RAG system with various options:
# Basic chunk listing
rlama list-chunks [rag-name]
# With content preview (shows first 100 characters)
rlama list-chunks [rag-name] --show-content
# Filter by document name/ID substring
rlama list-chunks [rag-name] --document=readme
# Combine options
rlama list-chunks [rag-name] --document=api --show-contentOptions:
--show-content: Display chunk content preview--document: Filter by document name/ID substring
Output columns:
- Chunk ID (use with view-chunk command)
- Document Source
- Chunk Position (e.g., "2/5" for second of five chunks)
- Content Preview (if enabled)
- Created Date
Display detailed information about a specific chunk.
rlama view-chunk [rag-name] [chunk-id]Parameters:
rag-name: Name of the RAG systemchunk-id: Chunk identifier from list-chunks
Example:
rlama view-chunk documentation doc123_chunk_0Add new documents to an existing RAG system.
rlama add-docs [rag-name] [folder-path] [flags]Parameters:
rag-name: Name of the RAG systemfolder-path: Path to documents folder
Example:
rlama add-docs documentation ./new-docs --exclude-ext=.tmpAdd content from a website to an existing RAG system.
rlama crawl-add-docs [rag-name] [website-url]Parameters:
rag-name: Name of the RAG systemwebsite-url: URL of the website to crawl and add to the RAG
Options:
--max-depth: Maximum crawl depth (default: 2)--concurrency: Number of concurrent crawlers (default: 5)--exclude-path: Paths to exclude from crawling (comma-separated)--chunk-size: Character count per chunk (default: 1000)--chunk-overlap: Overlap between chunks in characters (default: 200)
Example:
# Add blog content to an existing RAG
rlama crawl-add-docs my-docs https://blog.example.com
# Customize crawling behavior
rlama crawl-add-docs knowledge-base https://docs.example.com --max-depth=1 --exclude-path=/apiUpdate the LLM model used by a RAG system.
rlama update-model [rag-name] [new-model]Parameters:
rag-name: Name of the RAG systemnew-model: New Ollama model name
Example:
rlama update-model documentation deepseek-r1:7b-instructChecks if a new version of RLAMA is available and installs it.
rlama update [--force/-f]Options:
--forceor-f: (Optional) Update without asking for confirmation.
Displays the current version of RLAMA.
rlama --versionor
rlama -vTo uninstall RLAMA:
If you installed via go install:
rlama uninstallRLAMA stores its data in ~/.rlama. To remove it:
rm -rf ~/.rlamaRLAMA supports many file formats:
- Text:
.txt,.md,.html,.json,.csv,.yaml,.yml,.xml,.org - Code:
.go,.py,.js,.java,.c,.cpp,.cxx,.h,.rb,.php,.rs,.swift,.kt,.ts,.tsx,.f,.F,.F90,.el,.svelte - Documents:
.pdf,.docx,.doc,.rtf,.odt,.pptx,.ppt,.xlsx,.xls,.epub
Installing dependencies via install_deps.sh is recommended to improve support for certain formats.
If you encounter connection errors to Ollama:
- Check that Ollama is running.
- By default, Ollama must be accessible at
http://localhost:11434or the host and port specified by the OLLAMA_HOST environment variable. - If your Ollama instance is running on a different host or port, use the
--hostand--portflags:rlama --host 192.168.1.100 --port 8000 list rlama --host my-ollama-server --port 11434 run my-rag
- Check Ollama logs for potential errors.
If you encounter problems with certain formats:
- Install dependencies via
./scripts/install_deps.sh. - Verify that your system has the required tools (
pdftotext,tesseract, etc.).
If the answers are not relevant:
- Check that the documents are properly indexed with
rlama list. - Make sure the content of the documents is properly extracted.
- Try rephrasing your question more precisely.
- Consider adjusting chunking parameters during RAG creation
For any other issues, please open an issue on the GitHub repository providing:
- The exact command used.
- The complete output of the command.
- Your operating system and architecture.
- The RLAMA version (
rlama --version).
RLAMA provides multiple ways to connect to your Ollama instance:
-
Command-line flags (highest priority):
rlama --host 192.168.1.100 --port 8080 run my-rag
-
Environment variable:
# Format: "host:port" or just "host" export OLLAMA_HOST=remote-server:8080 rlama run my-rag
-
Default values (used if no other method is specified):
- Host:
localhost - Port:
11434
- Host:
The precedence order is: command-line flags > environment variable > default values.
# Quick answers with minimal context
rlama run my-docs --context-size=10
# Deep analysis with maximum context
rlama run my-docs --context-size=50
# Balance between speed and depth
rlama run my-docs --context-size=30rlama rag llama3 my-project ./code \
--exclude-dir=node_modules,dist \
--process-ext=.go,.ts \
--exclude-ext=.spec.ts# List chunks with content preview
rlama list-chunks my-project --show-content
# Filter chunks from specific document
rlama list-chunks my-project --document=architectureGet full command help:
rlama --helpCommand-specific help:
rlama rag --help
rlama list-chunks --help
rlama update-model --helpAll commands support the global --host and --port flags for custom Ollama connections.
The precedence order is: command-line flags > environment variable > default values.
RLAMA now supports using GGUF models directly from Hugging Face through Ollama's native integration:
# Search for GGUF models on Hugging Face
rlama hf-browse "llama 3"
# Open browser with search results
rlama hf-browse mistral --openBefore creating a RAG, you can test a Hugging Face model directly:
# Try a model in chat mode
rlama run-hf bartowski/Llama-3.2-1B-Instruct-GGUF
# Specify quantization
rlama run-hf mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated-GGUF --quant Q5_K_MUse Hugging Face models when creating RAG systems:
# Create a RAG with a Hugging Face model
rlama rag hf.co/bartowski/Llama-3.2-1B-Instruct-GGUF my-rag ./docs
# Use specific quantization
rlama rag hf.co/mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated-GGUF:Q5_K_M my-rag ./docs