Deploy GPT Researcher

Deep research agent with cited reports, login gateway and private SearXNG

Deploy GPT Researcher

Just deployed

/etc/searxng

Deploy and Host GPT Researcher on Railway

GPT Researcher is an open-source deep research agent. You ask a question, and it plans sub-queries, searches the web, reads and filters the sources, and writes a long report with citations that you can export as PDF, Word or Markdown. It can also research documents you upload and answer follow-up questions about a report.

About Hosting GPT Researcher

GPT Researcher upstream ships a Python backend (FastAPI, with a websocket that streams each research run) and a Next.js web UI. It has no login of its own and fetches any URL it is given, so it should not face the internet directly. This template runs four services on Railway's private network: a login gateway (the only public service), the backend, the web UI and a private SearXNG search engine, so web search works without a paid search API. You supply one LLM key. Reports, research history and uploaded documents live on a volume and survive redeploys. Official 3.8.0 images are pinned by digest.

Common Use Cases

  • Market, competitor and product research: a cited report from many sources in a few minutes
  • Literature and technical reviews across papers, documentation and blog posts
  • Research over your own files: upload PDF, Word or text documents and combine them with web sources

Dependencies for GPT Researcher Hosting

  • An OpenAI API key (OPENAI_API_KEY), the only value the deploy form asks for. Other LLM providers work too (see below).
  • SearXNG, bundled as a private service, as the web search engine. No search API key is needed.

Deployment Dependencies

Implementation Details

Services:

  • GPT Researcher (public): a Caddy gateway. It asks for a username and password, then sends the API, websocket and file paths to Backend and everything else to Frontend, all on one origin. Health check /healthz.
  • Backend: the GPT Researcher API (port 8000). A volume at /data keeps reports.json (research history and chats), outputs/ (generated PDF, Word and Markdown files) and my-docs/ (uploaded documents).
  • Frontend: the Next.js web UI (port 3000), private.
  • SearXNG: private metasearch engine with JSON output enabled, used as the default retriever. Its settings.yml is kept on a volume at /etc/searxng.

After deploy:

  1. Open the GPT Researcher service's variables and copy UI_USERNAME (default admin) and the generated UI_PASSWORD.
  2. Open the public domain, sign in and ask your first question.

To use another LLM provider, set FAST_LLM, SMART_LLM, STRATEGIC_LLM and EMBEDDING (format provider:model, for example anthropic:claude-sonnet-4-5) and that provider's key on Backend, or set OPENAI_BASE_URL for an OpenAI-compatible endpoint. To use Tavily or another search API instead of SearXNG, set RETRIEVER and its key on Backend.

Anyone with the gateway password can fully control the backend, including its MCP settings, which can run commands inside the container. Share it only with people you trust, and don't give Backend or Frontend a public domain.

Updates: the services build from the template repository. When it moves to a newer GPT Researcher or SearXNG release, Railway shows an update notice and you choose when to apply it. Your data stays on the volumes. The repository's CHANGELOG.md lists breaking changes first.

Why Deploy GPT Researcher on Railway?

Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.

By deploying GPT Researcher on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway.


Template Content

More templates in this category

View Template
Chat Chat
Chat Chat, your own unified chat and search to AI platform.

okisdev
117
View Template
Weaviate
Vector database with hybrid search and an API key-protected API

GitHub Locker
1
View Template
Möbius Personal AI OS
Self-hosted AI agent that builds and runs apps for you, with memory.

bento
1