> In order to start Fredy, you must provide a config.json. As a start, use the one in this repo: https://github.com/orangecoding/fredy/blob/master/conf/config.json
Starting with **V20**, Fredy ships with a built-in **MCP Server **. This allows you to connect Fredy to LLMs (like Claude, ChatGPT, or local models via LM Studio) and query your real estate data using natural language.
The local LLM can even enrich existing listings by checking the listing online.
Immoscout has implemented advanced bot detection. In order to work around this, we are using a reversed engineered version of their mobile api. See [Immoscout Reverse Engineering Documentation](https://github.com/orangecoding/fredy/blob/master/reverse-engineered-immoscout.md)
Most browser-based providers (immowelt, immonet, kleinanzeigen, ...) are scraped through a hardened headless browser ([CloakBrowser](https://www.npmjs.com/package/cloakbrowser)). It makes the **browser fingerprint** indistinguishable from a real Chrome, which is enough when you run Fredy on a normal home connection.
On a **server / VPS the requests usually originate from a datacenter IP**, and providers behind anti-bot systems (e.g. AWS CloudFront/WAF) block those based on **IP reputation alone**, no matter how perfect the fingerprint is. The typical symptom: it works locally but you get `We have been detected as a bot :-/` on the server.
### The fix: a residential proxy
A **residential proxy** routes Fredy's browser through the internet connection of a real household, so the provider sees a "normal user" IP instead of a datacenter. For German portals, use a **German (DE) residential** (or mobile/4G) proxy. Plain VPNs and **datacenter proxies do not help** here, they share the same bad reputation as your server.
Leave the field empty to disable. The proxy applies to all headless-browser providers and takes effect on the next job run (no restart needed). Immoscout uses a separate mobile API and is not affected.
### Where to get a residential proxy
Residential proxies are a paid service (usually billed per GB, Fredy's traffic is small). Well-known providers offering German residential IPs include:
| Provider | Notes |
|---|---|
| [IPRoyal](https://iproyal.com) | Pay-as-you-go, no monthly minimum, good for low volume |
| [Webshare](https://www.webshare.io) | Cheap entry tier, has a small free plan to test with |
This is not an endorsement, pick whatever fits your budget. For low-volume use like Fredy, a pay-as-you-go plan (e.g. IPRoyal) or a cheap entry tier (e.g. Webshare) is usually plenty. Make sure to select **Germany** as the proxy location and keep the search interval reasonable (the higher the interval, the less you look like a bot).
The data includes: names of active adapters/providers, OS, architecture, Node version, and language. The information is entirely anonymous and helps me understand which adapters/providers are most frequently used.</p>
Fredy's UI is fully multilingual. Translation files live in `ui/src/locales/`. To add a new language, create a single JSON file there, no code changes required.
**Example: `ui/src/locales/fr.json`**
```json
{
"_meta": {
"flag": "🇫🇷",
"name": "Français",
"locale": "fr-FR",
"semiLocale": "fr"
},
"nav.dashboard": "Tableau de bord",
"common.save": "Enregistrer",
...
}
```
The `_meta` fields:
| Field | Description |
|---|---|
| `flag` | Unicode flag emoji shown in the language selector |
| `name` | Display name shown in the language selector |
| `locale` | BCP 47 locale string used for date and number formatting (e.g. `fr-FR`) |
> **Important:** `semiLocale` must exactly match a locale filename from the Semi UI locale sources (without the `.js` extension). See the [available Semi UI locales on GitHub](https://github.com/DouyinFE/semi-design/tree/main/packages/semi-ui/locale/source) for the full list of supported keys.
After adding the file, rebuild the frontend (`yarn build:frontend` or restart the dev server) and the new language will appear automatically in **Settings → User Settings → Language**.
When I started building Fredy, LLMs were still basically the wet dream of a few nerdy scientists.
Nowadays, it’s easier than ever to throw a prompt into the LLM of your choice and let 'the AI' build your stuff. I’m not against that. I use Claude Code myself for smaller tasks, and I do think these tools can be really useful.
That said, I still believe humans should stay in charge. AI is great-ish at writing code, but it still lacks creativity, context, and the ability to see the full picture.
So, if you want to contribute to Fredy, using AI tools to get things done is totally fine. Just please don’t stop thinking.
I’ve had one too many PRs full of hallucinated bullshit.
