Deploy Perplexica — A Privacy-Focused AI Search Engine — to Your Own Server

Perplexica is a privacy-focused AI answering engine that runs entirely on your own server. It combines web search via SearxNG with local LLMs (Ollama) or cloud providers (OpenAI, Claude, Gemini, Groq) to deliver accurate answers with cited sources — all while keeping your searches completely private.

Setting it up manually involves configuring Docker, SearxNG, and connecting AI providers. With HowToDeploy, the entire stack is provisioned and running in under a minute.

Why Perplexica?

• Privacy-first — all searches stay on your server, nothing is tracked or sold
• Multiple AI providers — Ollama, OpenAI, Claude, Gemini, Groq, or any OpenAI-compatible endpoint
• Smart search modes — Speed Mode for quick answers, Balanced for everyday use, Quality for deep research
• Cited sources — every answer includes references so you can verify the information
• Image and video search — find visual content alongside text results
• File uploads — ask questions about PDFs, text files, and images
• Academic search — search research papers and scholarly sources
• Search history — every query is saved locally for easy reference
• Open source — MIT licensed, 32k+ GitHub stars, active community

Prerequisites

Before you start, you'll need:

• A HowToDeploy account (sign up free)
• A cloud provider API key (DigitalOcean, Hetzner, Vultr, Linode, or AWS)

Step 1: Connect your cloud provider

Go to Settings → Cloud Providers and paste your API key.

Tip: Perplexica runs well on a 4GB RAM / 2 CPU server. If you plan to run local LLMs with Ollama on the same server, consider a larger instance (8GB+ RAM).

Step 2: Deploy Perplexica

Head to the Dashboard and find Perplexica in the AI Agents section. Click the card to open the deploy form.

No required fields — Perplexica has a built-in setup screen where you configure your AI providers after deployment. Server size, region, Docker, and SearxNG are all pre-configured.

Click Deploy and watch the progress in real time.

Step 3: Configure your AI provider

Once deployment completes, open the Perplexica URL. The setup screen lets you configure:

• OpenAI — paste your API key and select a model (GPT-4o, GPT-4, etc.)
• Ollama — point to a running Ollama instance for fully local inference
• Claude — use Anthropic's models for high-quality reasoning
• Gemini — connect Google's models
• Groq — fast inference with Groq's LPU hardware

You can also pre-fill these during deployment by expanding Advanced Settings and entering your OpenAI API key or Ollama URL.

Step 4: Add a custom domain (optional)

Want your search engine at search.yourdomain.com? Enter your domain in Advanced Settings before deploying.

After deployment, point an A record for your domain to the server IP, click Verify DNS, and Caddy issues the SSL certificate automatically.

Step 5: Start searching

Open Perplexica and try queries like:

• "What are the best open-source alternatives to Notion?"
• "Explain how transformers work in machine learning"
• "Find recent research papers on CRISPR gene editing"
• "Compare Next.js and Remix for building web apps"

Every answer includes cited sources you can click to verify. Switch between search modes depending on whether you need a quick answer or a deep dive.

Search modes explained

What's included

Every Perplexica deployment includes:

• Docker — containerised services for isolation and reproducibility
• SearxNG — bundled privacy-respecting meta search engine
• Perplexica — the full AI answering engine with web UI
• Caddy — automatic HTTPS with Let's Encrypt (when using a custom domain)
• Full SSH access — your server, your searches, your data

Perplexica vs. Perplexity AI

Choose Perplexica if: you want Perplexity-level answers with complete privacy, local LLM support, and no usage limits.

Pricing

You pay your cloud provider directly for the server (typically $6-24/month depending on size). HowToDeploy charges a small monthly management fee for monitoring and support.

Start with a 7-day free trial — no credit card required.

---

Ready to run your own private AI search engine? Deploy Perplexica now →