Open RAGFlow

Open-source RAG engine fusing RAG with Agent capabilities. Full-stack: Python backend (Flask), React/TypeScript frontend, Docker-deployed microservices.

Use Cases

• Deploying or self-hosting RAGFlow via Docker Compose or from source
• Configuring RAGFlow (LLM providers, API keys, document engines, ports)
• Managing knowledge bases, datasets, documents, agents, and chats via RAGFlow CLI
• Understanding RAGFlow architecture (DeepDoc, Agent system, RAG pipeline)
• Integrating with RAGFlow REST API
• Troubleshooting RAGFlow deployment or runtime issues.

Quick Reference

• Website: https://ragflow.io
• Cloud: https://cloud.ragflow.io
• Docs: https://ragflow.io/docs/dev/
• Repo: https://github.com/infiniflow/ragflow
• Discord: https://discord.gg/NjYzJD3GM3
• Docker Hub: infiniflow/ragflow
• License: Apache 2.0

When to Use Which Reference

• Deploying / troubleshooting deployment → references/deployment.md
• Understanding architecture / components / data flow → references/architecture.md
• Using CLI to manage datasets, agents, models → references/cli-reference.md

Prerequisites

• CPU >= 4 cores, RAM >= 16 GB, Disk >= 50 GB
• Docker >= 24.0.0 & Docker Compose >= v2.26.1
• vm.max_map_count >= 262144 (Linux, for Elasticsearch)
• gVisor: optional, only needed for the code executor (sandbox) feature

Docker Deployment (Quick Start)

git clone https://github.com/infiniflow/ragflow.git
cd ragflow/docker
docker compose -f docker-compose.yml up -d
docker logs -f docker-ragflow-cpu-1  # wait for the banner, then login
# Open http://YOUR_SERVER_IP in browser

Configure LLM API keys in docker/service_conf.yaml.template under user_default_llm, then restart:

docker compose -f docker-compose.yml up -d

Chinese Mirror Images

If Docker Hub is slow:

• Huawei Cloud: swr.cn-north-4.myhuaweicloud.com/infiniflow/ragflow
• Alibaba Cloud: registry.cn-hangzhou.aliyuncs.com/infiniflow/ragflow

Set HuggingFace mirror if needed: HF_ENDPOINT=https://hf-mirror.com in docker/.env.

Key Configuration Files

Post-Deployment LLM Setup

1. Log in via browser
2. Edit docker/service_conf.yaml.template:

   user_default_llm:
     factory: "OpenAI"   # or "DeepSeek", "Gemini", etc.
     api_key: "sk-..."
     base_url: "https://api.openai.com/v1/"
   

3. Run docker compose -f docker-compose.yml up -d to apply.

CLI Quick Reference

All CLI commands end with ;. Full reference: references/cli-reference.md.

# Datasets
LIST DATASETS;
CREATE DATASET 'my_kb' WITH EMBEDDING 'text-embedding-ada-002' PARSER 'pdf';
DROP DATASET 'my_kb';
LIST FILES OF DATASET 'my_kb';

# Documents
IMPORT '/path/to/doc.pdf' INTO DATASET 'my_kb';
PARSE DATASET 'my_kb' SYNC;
PARSE DATASET 'my_kb' ASYNC;

# Search
SEARCH 'What is RAG?' ON DATASETS 'my_kb';

# Models
CREATE MODEL PROVIDER 'openai' 'sk-...';
SET DEFAULT LLM 'gpt-4';
LIST MODEL PROVIDERS;
LIST DEFAULT MODELS;

# Agents & Chats
LIST AGENTS;
LIST CHATS;
CREATE CHAT 'my_session';
DROP CHAT 'my_session';

# Connection
PING;
SHOW CURRENT USER;

Switching Doc Engine (Elasticsearch → Infinity)

docker compose -f docker/docker-compose.yml down -v  # WARNING: clears data
# Edit docker/.env: set DOC_ENGINE=infinity
docker compose -f docker-compose.yml up -d

Infinity is lighter weight but Linux/arm64 is not officially supported.

Architecture at a Glance

Web UI (React+TS+vitejs+shadcn) → Flask API (/api/) → RAG Core (/rag/) + Agent (/agent/)
                                                          ↓
Infrastructure: MySQL + Elasticsearch/Infinity + Redis + MinIO

• Backend (/api/): Flask blueprints — kb, dialog, document, canvas, file, user
• RAG Core (/rag/): DeepDoc parsing, LLM/embedding/rerank abstractions, chunking, GraphRAG
• Agent (/agent/): Canvas-based workflow builder with components (LLM, Retrieval, Code Executor, MCP, Search, SQL)
• Frontend (/web/): React 18 + TypeScript + Vite

See references/architecture.md for detailed component breakdown.

Development from Source

git clone https://github.com/infiniflow/ragflow.git && cd ragflow
uv sync --python 3.12 && uv run python3 download_deps.py
docker compose -f docker/docker-compose-base.yml up -d
# Add to /etc/hosts: 127.0.0.1 es01 infinity mysql minio redis sandbox-executor-manager
source .venv/bin/activate && export PYTHONPATH=$(pwd)
bash docker/launch_backend_service.sh
# Separate terminal:
cd web && npm install && npm run dev

Troubleshooting Quick Reference

API & SDK

• REST API: http://SERVER_IP/api/ — Swagger docs at /api/docs
• Python SDK: available in sdk/python/
• CLI client: python admin/client/ragflow_cli.py <command>
• Guide: https://ragflow.io/docs/dev/