OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
RAG & Knowledge Systems

Rag Platform

RAG Platform is an open-source Python system for ingesting knowledge, indexing multimodal content and serving citation-ready retrieval through hybrid search, tenant-isolated APIs and MCP tools.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Rag Platform — OpenViST project profile in RAG & Knowledge Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when building and operating a self-hosted knowledge retrieval backend for applications or AI agents, especially when tenant isolation, hybrid retrieval and citation metadata matter.

Who it is for

  • Developers building knowledge search into applications.
  • Teams creating retrieval backends for AI agents.
  • Operators who can provide the required databases, vector store and embedding endpoint.

Documented uses

  • Connect Yuque or local Markdown sources and process documents through download, cleaning, chunking and vectorization.
  • Search indexed knowledge using vector similarity, BM25 and title relevance.
  • Serve tenant-isolated ingestion, search, update and deletion operations through a versioned API.
  • Return retrieval metadata and source details that applications can use for citations.
  • Expose retrieval and health checks to model-facing clients through MCP.
Before you choose

Requirements & limits

The public repository identifies the project as open source under the MIT license. Self-hosting requires the documented infrastructure and an external compatible embedding endpoint.

  • Python 3.10+, uv, PostgreSQL 14+, Qdrant and a compatible embedding endpoint are listed requirements; Redis is needed for rag-search.
  • Docker and Docker Compose are needed for the documented container workflow.
  • The included connectors are Yuque and local Markdown; other source integrations depend on connectors.
  • PDF, Office, OCR and legacy document extraction may need additional system utilities.
  • The README marks packages as version 0.1.0 and says public APIs may change before a stable release.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install RAG Platform and prepare local infrastructure

Clone the repository, install its workspace packages, configure credentials, and start the local PostgreSQL, Qdrant, and Redis services.

Before you start

  • Python 3.10 or newer
  • uv
  • Docker and Docker Compose
  • PostgreSQL, Qdrant, Redis, and a compatible external embedding endpoint; the Compose file provides the first three services

Steps

  1. Clone the repository and enter its directory: the command below then cd rag-platform.

    Documented command
    git clone https://github.com/Today-Hbw/rag-platform.git

  2. Install workspace dependencies and create the environment file: the command below and cp .env.example .env (PowerShell: Copy-Item .env.example .env).

    Documented command
    uv sync --all-extras --dev

  3. Edit .env with PostgreSQL, Qdrant, Redis, embedding, and API credentials; do not commit this file.

  4. Start local infrastructure with the command below. The Compose PostgreSQL service applies migrations when initializing a new postgres_data volume; for an external PostgreSQL instance, apply 0001_init.sql, 0003_rag_collection.sql, then `0004_p

    Documented command
    docker compose -f deploy/docker-compose.yml --env-file .env up -d postgres qdrant redis

Expected outcome

Workspace dependencies are installed, credentials are configured locally, and the Compose-managed infrastructure is started. PostgreSQL migrations are initialized automatically for a new data volume or can be applied in order to an external instance.

Guide limitations

  • The embedding provider is an external dependency.
  • PostgreSQL initialization scripts run only when the data directory is empty.
  • PDF, Office, OCR, and legacy-document extraction may require additional system tools such as Poppler, Tesseract, LibreOffice, and Antiword.
  • The project is under active development; public APIs may evolve before the first stable release.
Documentation 1 ↗
Source-based practical guide

Run the Yuque ingestion and indexing pipeline

Preview Yuque source changes, then download, clean, and index content using the documented pipeline stages.

Before you start

  • A configured RAG Platform environment and compatible embedding endpoint
  • Yuque connector configuration
  • PostgreSQL and Qdrant services available to the pipeline

Steps

  1. Optionally preview source changes without writing files or updating the database: rag-pipeline sync --source yuque --dry-run.

  2. Download changed Yuque documents: rag-pipeline sync --source yuque.

  3. Convert downloaded content into normalized Markdown: rag-pipeline clean.

  4. Chunk, embed, and index the cleaned documents: rag-pipeline vectorize.

Expected outcome

The pipeline processes Yuque content through synchronization, cleaning, chunking, embedding, and indexing as separate stages that can be retried and operated independently.

Guide limitations

  • The embedding endpoint must be compatible and configured; vectorization relies on it.
  • The ingestion pipeline is documented as staged; this recipe does not establish a successful run for a particular environment or source configuration.
  • Public APIs may evolve while the project remains under active development.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
Today-Hbw/rag-platform · Official homepage title · retained in its source language
How we review sourcesSuggest a correction