fredy

Website  |   Demo  |   Quick Start  |   Docs

Fredy - Your Self-Hosted Real Estate Finder for Europe

Fredy watches 24 real estate portals across for you (e.g. Immoscout, Kleinanzeigen etc), drops duplicates across platforms, and notifies you via Slack, Telegram, Email, ntfy, Discord and more the moment a new listing appears. Searches are managed from a Web UI, and you never see the same listing twice.

On top of the listing itself, Fredy answers the three questions a portal will not:

• What would this cost me? Enter your income, living costs and savings once, and every listing is measured against them, for renting and for buying. → Financing
• How long would I actually travel? Not straight-line distance, but the real journey from your own address by public transport, car, bike or on foot. → Travel time
• Is this too good to be true? Fredy reads every listing for the marks of a rental scam and warns you when enough of them line up. → Scam detection

Quick Start

Docker (recommended)

docker run -d --name fredy \
  -v fredy_conf:/conf \
  -v fredy_db:/db \
  -p 9998:9998 \
  ghcr.io/orangecoding/fredy:master

Open http://localhost:9998 and sign in with admin / admin.

[!NOTE] No configuration file is needed to start. Fredy creates /conf/config.json on first run if it is missing. That file only holds the database path, everything else is configured in the Web UI and stored in the database. Documents you upload live in the database too, so the /db volume above is all you need to keep them, and every backup already contains them.

:master follows the master branch; pin a release with its version tag instead, e.g. ghcr.io/orangecoding/fredy:28.0.0. Images are built for linux/amd64 and linux/arm64. Logs: docker logs fredy -f.

There is also a pre-release channel following the develop branch, ghcr.io/orangecoding/fredy-pre-release:latest. It is built from the same pipeline (lint, format check and the offline test suite all have to pass first), but the changes in it have not been through master yet. Use it to try upcoming features or to verify a fix, not for an instance you rely on.

Node.js

Requires Node.js 22.22.0 or higher.

yarn
yarn run build:frontend  # builds the Web UI into ui/public
yarn run start:backend   # serves the UI and the API on port 9998

Unraid

Fredy is in the Unraid community store.

✨ What you get

• 24 portals across : ImmoScout24, Immowelt, Kleinanzeigen, WG-Gesucht,

willhaben, Flatfox, idealista, Subito and [16 more](doc/providers.md)

• ⚡ Instant notifications: Slack, Telegram, Email (SMTP, SendGrid, Mailjet, Resend), ntfy, Discord, Mattermost, Pushover, Apprise and more
• Deduplication across platforms: the same flat advertised on ImmoScout, Immowelt and

Kleinanzeigen reaches you once, matched on living space, rooms and location rather than on the
headline, no two portals write that the same way

• Scam detection: a warning on listings that read like rental fraud,

with your verdict always outranking Fredy's

• Financing calculator: which listings you can afford, for renting and

for buying

• Real travel times from your addresses by public transport, car, bike

or on foot, plus a filter to match, and the transport network with live departures on the map

• Market benchmark: the price per m² of every listing against the local median, so cheap and

expensive are facts rather than a feeling

• Document uploads: attach the exposé, floor plans and photos to a listing, so they survive

the portal taking the ad down. A listing with documents is never cleaned up automatically

• ️ Web UI in several languages, with configurable search intervals and working hours
• MCP server: query your listings from Claude, ChatGPT or a local LLM
• Runs anywhere: Docker, Node.js, self-hosted, and uses the reverse engineered

**ImmoScout Mobile API**

Basic architecture overview

A job can hold as many channels as you want, including several of the same type, and every new listing goes out through all of them at once. Rotating a token means editing one channel instead of every job that used it. Channels belong to whoever created them; an admin can share one with all users or with admins only, which lets others send through it without ever revealing its credentials. A channel still used by a job cannot be deleted.

…

Going deeper

MCP Server

Since V20, Fredy ships a built-in MCP Server, so you can connect it to Claude, ChatGPT or a local model (LM Studio) and query your real estate data in natural language. A local LLM can even enrich existing listings by checking them online. Setup: MCP Readme.

To connect Claude.ai or ChatGPT over OAuth, set Fredy's baseUrl to its public HTTPS URL and add <baseUrl>/api/mcp as a custom MCP server. Fredy advertises OAuth discovery metadata, dynamically registers the client, and asks you to sign in and approve read access. Access tokens expire after one hour and refresh automatically; existing MCP tokens keep working for local clients. Connected apps are listed under Settings → Connections, where access can be revoked at any time.

Analytics

Fredy is completely free (and will always remain free). However, it would be a huge help if you would allow me to collect some analytical data. Before you freak out, let me explain: if you agree, Fredy sends a ping every 6 hours to my internal tracking project (will be open sourced soon). The data includes which notification adapters and providers are in use (the type only, for example slack, never your channels, their names or their credentials), OS, architecture, Node version and language. It is entirely anonymous and helps me understand which adapters and providers matter most.

Thanks

Sponsorship

I build and maintain Fredy in my free time. If it saves you some, consider chipping in ❤️

Fredy is backed by the JetBrains Open Source Support Program.

Contributing

Thanks to everyone who has contributed!

Start with the Contributing Guide; it also covers where AI-assisted contributions help, and where they do not.

️ Data Attribution & Partners

Timetables, journey planning and travel times come from Transitous, a community-run MOTIS instance. It is free, needs no API key, and is maintained by volunteers, so please be considerate with the load you put on it and read their usage policy before pointing a large instance at it. Street and map data come from OpenStreetMap contributors.

Every listing with coordinates links out to Lagecheck, which rates the address on noise, air quality, green space and flood risk. Fredy only builds the link and stores nothing from it; the underlying data is by geosci.de.

License

Apache-2.0 with two additional conditions:

• Commons Clause - you