# AI use cases for Local Deep Research (LearningCircuit/local-deep-research)

Repo: [LearningCircuit/local-deep-research](https://github.com/LearningCircuit/local-deep-research) | Use cases: 3 | Page: https://dedupedai.com/use-cases/learningcircuit-local-deep-research/

Local Deep Research is an AI research assistant you run yourself. It performs deep, agentic research with any LLM and many search engines, cites its sources, builds a searchable library of what it downloads, and offers a web UI, a Python API and an MCP server.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Run a private research assistant with Docker Compose: https://dedupedai.com/use-cases/learningcircuit-local-deep-research/run-a-private-research-assistant-with-docker-compose/
- Run research from a Python script: https://dedupedai.com/use-cases/learningcircuit-local-deep-research/call-local-deep-research-from-python/
- Give your AI assistant deep research through MCP: https://dedupedai.com/use-cases/learningcircuit-local-deep-research/use-local-deep-research-as-an-mcp-server/

## How to run a private research assistant with Docker Compose with Local Deep Research

Repo: [LearningCircuit/local-deep-research](https://github.com/LearningCircuit/local-deep-research) | Difficulty: Beginner | Time: 30 minutes | Basis: From the README

Start Local Deep Research with its bundled Ollama and SearXNG using one docker compose command.

**Who it is for:** A researcher who wants cited reports without sending prompts to a hosted AI service.

**The problem:** Hosted research tools see every question you ask.

**The outcome:** A web UI on port 5000 that runs research with a local model and cites sources.

### Steps

1. Install Docker
2. Run `curl -O https://raw.githubusercontent.com/LearningCircuit/local-deep-research/main/docker-compose.yml && docker compose up -d`
3. Open the web UI and create your account
4. Ask a research question and read the cited report

### Prompt to paste into your agent

```text
Using LearningCircuit/local-deep-research, run `curl -O https://raw.githubusercontent.com/LearningCircuit/local-deep-research/main/docker-compose.yml && docker compose up -d`, then research <my question> and save the cited report.
```

### Caveat

A GPU makes local models much faster; there is a GPU override compose file.

Risk, privacy: A self hosted SearXNG still forwards search queries to upstream engines on the internet.

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/learningcircuit-local-deep-research/run-a-private-research-assistant-with-docker-compose/

## How to run research from a Python script with Local Deep Research

Repo: [LearningCircuit/local-deep-research](https://github.com/LearningCircuit/local-deep-research) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README

Install the pip package and use quick_query or LDRClient to get research summaries in code.

**Who it is for:** A developer who wants research summaries inside a pipeline.

**The problem:** Running research by hand in a UI does not scale to many questions.

**The outcome:** Research summaries returned as Python values for each question.

### Steps

1. Install with `pip install local-deep-research`
2. Start the server with `python -m local_deep_research.web.app` and create a user
3. Call quick_query with your username, password and question, as in the README
4. Or create an LDRClient, log in and call quick_research

### Prompt to paste into your agent

```text
Using LearningCircuit/local-deep-research, install `pip install local-deep-research`, start it with `python -m local_deep_research.web.app`, then write a script that runs quick_research for each question in <my list> and saves the summaries.
```

### Caveat

You need Ollama or another LLM endpoint and SearXNG running for full results.

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/learningcircuit-local-deep-research/call-local-deep-research-from-python/

## How to give your AI assistant deep research through MCP with Local Deep Research

Repo: [LearningCircuit/local-deep-research](https://github.com/LearningCircuit/local-deep-research) | Difficulty: Intermediate | Time: 20 minutes | Basis: From the README

Install the MCP extra and add ldr-mcp to your MCP client config with Ollama or OpenAI as the provider.

**Who it is for:** Users of MCP clients like Claude Desktop who want cited research tools.

**The problem:** Chat assistants answer from memory instead of searching arXiv or PubMed.

**The outcome:** An MCP server your assistant can call for search and research.

### Steps

1. Install with `pip install "local-deep-research[mcp]"`
2. Add an mcpServers entry with command ldr-mcp to your MCP client config
3. Set LDR_LLM_PROVIDER to ollama and LDR_LLM_OLLAMA_URL to your Ollama address, as in the README
4. Ask your assistant to search arXiv or PubMed through the tool

### Prompt to paste into your agent

```text
Using LearningCircuit/local-deep-research, install `pip install "local-deep-research[mcp]"`, add the ldr-mcp server to <my MCP client> config with the Ollama provider, and test a search on arXiv.
```

### Caveat

With a hosted provider such as OpenAI, your queries leave the machine.

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/learningcircuit-local-deep-research/use-local-deep-research-as-an-mcp-server/
