Deploy AIHOT

Self-running AI news site: collects, scores and writes daily briefings

Deploy AIHOT

/var/lib/postgresql/data

Just deployed

/data

Deploy and Host AIHOT on Railway

AIHOT is an open-source (MIT) news site that runs itself. It collects items from your sources, has a language model screen, score and summarise each one, groups reports of the same story into events ranked by how many sources cover them, and publishes a daily briefing plus weekly and monthly reports.

About Hosting AIHOT

AIHOT is a Node.js app with three processes: a server-rendered web site, an API and a background worker, all on PostgreSQL. This template runs them as two services: Web, which is public and holds no secrets, and Backend, which runs the API and the worker together because both use the same data volume for image caches, uploads and local backups. Backend applies the database migrations and adds the demo sources at every start. AIHOT publishes no image or release tags, so the template builds from a pinned upstream commit, checked by SHA-256. You supply one key for any OpenAI-compatible model API. Model calls are billed by your provider.

Common Use Cases

  • Run an AI news site that picks and summarises the day's stories by itself
  • Build a hot-topics site and daily briefing for another field by swapping in your own sources and selection rules
  • Give your team or readers RSS feeds, a public JSON API and an MCP endpoint for curated news

Dependencies for AIHOT Hosting

  • PostgreSQL 17 (included)
  • An API key for an OpenAI-compatible chat model (DeepSeek by default)

Deployment Dependencies

Implementation Details

Services:

  • Web: the public site and the admin at /admin. It reads everything from Backend over the private network.
  • Backend: the API and the worker (collection, model calls, grouping, reports, schedules), with a volume at /data. Private only.
  • Postgres: Railway's PostgreSQL 17 with its own volume. It matches the pg_dump the optional backups use.

The only required field is LLM_API_KEY. LLM_BASE_URL and LLM_MODEL default to DeepSeek. For another OpenAI-compatible provider, change both, and clear LLM_EXTRA_JSON if the provider rejects unknown fields.

After deploying:

  1. Open the Web service's URL. The admin is at /admin. Its password is ADMIN_PASSWORD in the Backend service's Variables tab.
  2. The demo sources start collecting within a couple of minutes. The first batch of about 150 items takes about half an hour of model calls.
  3. To make it about your own field, fork AIHOT, edit industry/ and site/ as its customising guide explains, and change the commit in the template repo's Dockerfiles. With your own domain, set SITE_URL on Web (Backend references it).

Optional features (embeddings, per-step models, paid collectors, S3 backups, Feishu, IndexNow) are Backend variables listed in upstream's .env.example.

Updates: the template builds from aalfath/aihot-railway-template. When we move the pinned commit, Railway shows an update notice. Its CHANGELOG.md lists breaking changes first. Backend runs new migrations by itself.

Why Deploy AIHOT 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 AIHOT 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.

Tracklayer
1