# AI use cases for Ragflow (infiniflow/ragflow) Repo: [infiniflow/ragflow](https://github.com/infiniflow/ragflow) | Use cases: 3 | Page: https://dedupedai.com/use-cases/infiniflow-ragflow/ RAGFlow is an open-source Retrieval-Augmented Generation engine that fuses RAG with agent capabilities to build a context layer for LLMs. It offers deep document understanding of complex formats, template-based chunking you can inspect, Knowledge Compilation into Wikis, Graphs, Trees and Mind Maps, Agentic Retrieval with Low to Ultra thinking modes, and grounded answers with traceable citations, deployed with Docker or from source. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Self-host a chat-with-your-documents engine: https://dedupedai.com/use-cases/infiniflow-ragflow/self-host-a-chat-with-your-documents-engine/ - See and fix how your documents are chunked: https://dedupedai.com/use-cases/infiniflow-ragflow/check-and-fix-how-your-documents-are-chunked/ - Answer multi-step questions with agentic retrieval: https://dedupedai.com/use-cases/infiniflow-ragflow/answer-hard-questions-with-agentic-retrieval/ ## How to self-host a chat-with-your-documents engine with Ragflow Repo: [infiniflow/ragflow](https://github.com/infiniflow/ragflow) | Difficulty: Intermediate | Time: 1 hour | Basis: From the README Start RAGFlow with Docker Compose, connect your models and ask questions over your own files. **Who it is for:** A small team that wants a private knowledge assistant over policies, manuals and reports. **The problem:** Hosted RAG tools mean uploading internal documents to someone else, and building your own pipeline takes weeks. **The outcome:** A RAGFlow web UI on your own server where you upload documents and get answers with citations. ### Steps 1. Clone the repo with `git clone https://github.com/infiniflow/ragflow.git` and enter the folder with `cd ragflow/docker` 2. Check out the release tag with `git checkout v1.0.0-rc3` 3. Start the services with `docker compose -f docker-compose.yml up -d` 4. Confirm it is ready with `curl -f http://localhost/api/v1/system/healthz`, and check logs with `docker logs --tail 50 ragflow-cpu` if it is not 5. Open the server IP in a browser, sign in, and add an LLM, embedding model and reranker on the model provider page ### Prompt to paste into your agent ```text Deploy infiniflow/ragflow on this server with Docker Compose at tag v1.0.0-rc3, wait until the healthz endpoint returns 200, and tell me the URL to open. List what I still need to configure on the model provider page. ``` ### Caveat In the 1.0 open-source release DeepDoc uses CPU inference for layout, OCR and tables, so parsing large document sets takes time. Risk, credentials: You paste model provider API keys into RAGFlow, so restrict access to the server and change default passwords in docker/.env before exposing it. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/infiniflow-ragflow/self-host-a-chat-with-your-documents-engine/ ## How to see and fix how your documents are chunked with Ragflow Repo: [infiniflow/ragflow](https://github.com/infiniflow/ragflow) | Difficulty: Beginner | Time: 30 minutes | Basis: From the README Use RAGFlow's chunk visualization and templates so retrieval finds the right passages. **Who it is for:** Someone whose RAG answers keep missing facts that are clearly in the source PDF. **The problem:** Most RAG tools chunk blindly, so a table or clause gets split and the right passage never comes back. **The outcome:** Documents chunked with a template that fits them, with chunks you have inspected and corrected by hand. ### Steps 1. Deploy RAGFlow with Docker as in the README and sign in 2. Upload a document and pick a chunking template that fits its type 3. Open the chunk visualization and review where the text was split 4. Adjust chunks by hand where needed; the README calls this human intervention 5. Ask a question and click through the traceable citations to confirm the right chunk was used ### Prompt to paste into your agent ```text In my infiniflow/ragflow instance, upload , try two chunking templates, and compare how each splits
. Recommend the template that keeps
intact and explain why. ``` ### Caveat Template choice is a judgement call per document type; the README lists the feature but not which template suits which file. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/infiniflow-ragflow/check-and-fix-how-your-documents-are-chunked/ ## How to answer multi-step questions with agentic retrieval with Ragflow Repo: [infiniflow/ragflow](https://github.com/infiniflow/ragflow) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Turn on Agentic Retrieval so the model breaks down a hard question, retrieves several times and checks evidence. **Who it is for:** An analyst asking questions that span several documents, such as comparing terms across contracts. **The problem:** Single-shot retrieval returns a few chunks and misses context spread across many files. **The outcome:** Better grounded answers built from multiple rounds of retrieval, at a thinking depth you choose. ### Steps 1. Deploy RAGFlow and load the documents into a dataset 2. Enable Agentic Retrieval for complex questions 3. Pick a thinking mode: Low, Medium, High or Ultra, depending on how complex the question is 4. Optionally use Knowledge Compilation to generate a Wiki, Graph or Timeline of the dataset first 5. Check the citations in the answer before relying on it ### Prompt to paste into your agent ```text In infiniflow/ragflow, enable Agentic Retrieval on dataset and ask at Medium and then High thinking mode. Compare the two answers and their citations and tell me which is better supported. ``` ### Caveat RAGFlow 1.0 is at release candidate stage, so expect changes between versions. Risk, cost: Higher thinking modes make more model calls per question, which costs more with a paid LLM provider. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/infiniflow-ragflow/answer-hard-questions-with-agentic-retrieval/