mcp-registry/mcp-server-perplexity

mcp-server-perplexity

Author: tanigami

Description: Python-based MCP (Model Context Protocol) server that exposes Perplexity AI’s chat-completion API as an MCP tool (ask_perplexity).

Stars: 76

Forks: 30

License: MIT License

Category: Specialized

Overview

1. Prerequisites

• Python 3.9+

• git

2. Clone the repository

   git clone https://github.com/tanigami/mcp-server-perplexity.git
   cd mcp-server-perplexity

3. Create and activate a virtual environment (recommended)

   python -m venv .venv
   source .venv/bin/activate   # Windows: .venv\Scripts\activate

4. Install Python dependencies

pip install -r requirements.txt

5. Configure credentials

The server needs a Perplexity API token (and optionally an OpenAI key if the tool chains to OpenAI):

   export PERPLEXITY_API_KEY="<your-perplexity-key>"
   # optional
   export OPENAI_API_KEY="<openai-key>"

You can alternatively copy `.env.example` to `.env` and edit the values.

6. Run the MCP server locally

python -m mcp_server_perplexity --host 0.0.0.0 --port 8000

7. Verify

Open http://localhost:8000/health or use `curl`:

curl http://localhost:8000/health

You should receive `{ "status": "ok" }`.

8. Docker (optional)

   docker build -t mcp-perplexity .
   docker run -d -p 8000:8000 -e PERPLEXITY_API_KEY=<key> mcp-perplexity

Request chat completion with citations from Perplexity

License: MIT License

Updated 7/30/2025