Skip to content
QuickRAG
Menu

Document knowledge, ready to use

Your knowledge.Put to work.

Turn PDFs, manuals, and everyday files into searchable knowledge. Give your team a place to organize documents, and your applications a way to find the right source material.

Start self-hostingExplore the product

Source availableYour infrastructureYour model choice

Inside QuickRAGIllustrative workspace
Acme / KnowledgeIndexed

Retrieval test

HYBRID

What is the return policy?

Matching source passages

returns-policy.pdf

Returns within 30 days are accepted when the item is unused and in its original packaging.

Page 4 · chunk 07

support-handbook.docx

To start a return, ask for the order number and confirm the delivery date.

Section 2 · chunk 03

01Organize files

02Find context

03Trace sources

A sample query with illustrative documents. Retrieval returns source passages your application can use.

Built on tools your team already knows.

  • PostgreSQL + pgvector
  • Rust
  • Python
  • REST / MCP
  • Docker Compose

For the people behind the knowledge

One knowledge base. More useful everywhere.

Important information is scattered across files, folders, and systems. QuickRAG brings those documents into a shared retrieval layer so teams can maintain the knowledge and developers can build on it.

01 / Support

Keep answers close to the source.

Organize policies, product guides, and support handbooks. Give your support tools relevant passages to use when helping customers.

Policies · Help articles · Product manuals

02 / Operations

Make team knowledge easier to find.

Bring procedures and internal documentation into named knowledge bases. Track processing, keep documents current, and see where a result came from.

Runbooks · Procedures · Internal guides

03 / Engineering

Connect your product to your documents.

Retrieve context through an API instead of building an ingestion pipeline from scratch. Keep your application, models, and interface under your control.

Search tools · Assistants · Internal apps

RAG, in plain language: find relevant source text before an application writes an answer. QuickRAG handles finding that text; you choose how to use it.

From files to useful context

A clear path from document to result.

Manage the pipeline in a browser. Use the results wherever your team works.

  1. 01

    Bring your documents

    Upload files or connect ingestion sources. Group information into knowledge bases that fit your team.

    PDF · DOCX · CSV · Markdown · more

  2. 02

    Prepare the knowledge

    Extract text and split it into focused passages. Choose a chunking template that suits the material.

    Parsing · Chunking · Background jobs

  3. 03

    Choose your models

    Connect an embedding provider to make meaning searchable. Configure retrieval and optional reranking.

    Your providers · PostgreSQL + pgvector

  4. 04

    Retrieve with context

    Search from the console or your application. Get ranked passages with their document and chunk references.

    Semantic · Full-text · Hybrid

Built to fit your stack

Your interface. Your models. One retrieval layer.

Use QuickRAG behind a search experience, a support tool, or an assistant you build. REST exposes configurable retrieval; the JSON-RPC endpoint provides hybrid search for integrations.

Search your way
Combine vector similarity and full-text search. Adjust result count, score thresholds, filters, and reranking through REST.
Keep the source attached
REST results include passage content, score, filename, document_id, and chunk_id. Page references are included when available.
Use your own providers
Configure embedding and reranking providers. Generation happens in your application, with the model you choose.
Explore the API source
RESTPOST /api/v1/retrieval/search
curl -X POST "$QUICKRAG_URL/api/v1/retrieval/search" \
  -H "Authorization: Bearer $QUICKRAG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "What is the return policy?",
    "search_mode": "hybrid",
    "top_k": 5
  }'
See the MCP / JSON-RPC request
MCPHTTP / JSON-RPC
POST /api/v1/mcp
Authorization: Bearer <YOUR_API_KEY>
Content-Type: application/json

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "retrieval/search",
  "params": {
    "query": "What is the return policy?",
    "top_k": 5
  }
}

Set QUICKRAG_URL to your deployment and use a valid API key. The MCP endpoint accepts retrieval/search and uses hybrid retrieval; enable it with MCP_ENABLED.

A starting point for every team

Self-host today. Plan for what comes next.

The Community edition is available now. Team and Scale are proposed managed plans with indicative USD pricing; they are not available to purchase yet.

Available now · Self-hosted

Community

For builders and teams who want full control.

$0software

You pay for infrastructure and model usage.

  • Full source code
  • Knowledge bases and document ingestion
  • REST and JSON-RPC retrieval
  • Your infrastructure and model providers
Deploy Community

Proposed · Managed

Team

A proposed home for a growing team's knowledge.

$49USD / month

Indicative price, subject to launch details.

  • Community capabilities
  • Managed deployment, planned
  • Automated backups, planned
  • Workspace setup support, planned
Follow managed plans

Proposed · Managed

Scale

A proposed managed option for larger workloads.

$149USD / month

Indicative price, subject to launch details.

  • Proposed Team capabilities
  • Dedicated deployment options, planned
  • Capacity planning, planned
  • Deployment assistance, planned
Follow managed plans

Self-hosting has no QuickRAG subscription fee. Hosting, storage, and external model costs are separate. Proposed plan features, limits, and support terms will be defined before launch; there is no billing or checkout in this release.

A few practical answers

Before you get started.

Do I need to be a developer to use QuickRAG?

Once an administrator deploys and configures it, you can manage knowledge bases, upload documents, monitor processing, and test retrieval in your browser. Deployment and integration into other applications need technical setup.

Does QuickRAG include a chatbot or a language model?

QuickRAG prepares documents and retrieves relevant passages. It does not include a chatbot or a generation model. Connect your own embedding provider, then use the results in the interface or application you choose.

Where do my documents and data live?

With Community, you run QuickRAG on your own infrastructure. PostgreSQL stores knowledge and vectors, and S3-compatible storage holds files. Embedding and reranking providers may receive document text or queries; review their data policies when configuring them.

Can I sign up for Team or Scale now?

Not yet. Those are proposed managed offerings, with prices shown for planning. You can deploy Community today and follow the repository for managed plan updates.

Source available. Yours to run.

Build on knowledge you already have.

Deploy QuickRAG on your infrastructure, configure a provider, and start with one knowledge base. The deployment guide covers the stack, secrets, domains, and first-run checks.

Deploy with DokployView on GitHub
Get the code
git clone https://github.com/rifqiahmadf/quickrag.git
cd quickrag

Next: configure environment variables and provider credentials, then follow the deployment guide. A clone alone is not a running deployment.