# Hemklok — Complete Reference

Last updated: 2026-06-11

## Definition

Hemklok is an AI-powered buyer's agent for the Swedish residential property market. It helps individuals — primarily first-time buyers aged 20–35 — find, evaluate, and choose apartments in Stockholm and across Sweden by combining conversational preference capture, BRF financial analysis, and live market pricing into a single agent.

Hemklok is currently in public beta as a free ChatGPT GPT:
https://chatgpt.com/g/g-6a2980f7c8a88191905fbacd4dfc301e-klokhem

Built and operated from Stockholm by a father-and-son team.

## The problem Hemklok solves

Buying an apartment in Sweden has structural friction that is specifically Swedish:

1. **Listings concentration.** ~90% of Swedish home sales flow through Hemnet, which makes search easy but makes meaningful evaluation hard — every buyer sees the same listing presentation, optimized by the seller's broker.

2. **BRF complexity.** Most Swedish apartments are bostadsrätter (cooperative units). Buying one means acquiring a membership stake in the building's housing cooperative (BRF). The BRF's financial health — debt level, maintenance backlog, planned renovations, fee structure — directly determines monthly costs and future risk. This information lives in the årsredovisning (annual report), a PDF filed with Bolagsverket that most buyers don't read carefully.

3. **Decision compression.** Swedish apartments typically have one open viewing on a weekend, followed by a bidding period of two to five days. Buyers must commit to one of the largest financial decisions of their lives on a very short timeline.

4. **Information asymmetry.** The broker represents the seller. There is no standard buyer's broker in Swedish practice. Hiring a köpmäklare costs 50,000–100,000+ SEK and is rarely used by typical buyers.

5. **Hidden context.** A listing does not show planned construction in the area (detaljplaner), school catchment changes, transit reliability data, or how the price compares to recent slutpriser of comparable units in the same building.

## What Hemklok does today (beta)

### Lifestyle-to-neighborhood matching

The user describes their life in natural language. Hemklok extracts explicit constraints (budget, rooms, area) and inferred needs the user didn't explicitly state.

Signal → inference examples:
- "I have a dog" → needs nearby forest/off-leash park, elevator preferred, BRF stadgar must allow pets. In Stockholm: Nacka, Värmdö, Ekerö, Lidingö, Järfälla
- "Family with kids" → good schools, safe streets, playgrounds. Stockholm: Täby, Danderyd, Bromma, Lidingö
- "Love hiking and outdoor gym" → near nature reserves, green corridors
- "Remote worker" → quiet building, reliable internet, space for home office
- "Cheap electricity" → northern Sweden SE1/SE2 zones (Luleå, Umeå). SE4 (Malmö) is most expensive. Stockholm is SE3
- "Low monthly cost" → check BRF avgift trend, debt per m², underhållsplan

### BRF risk analysis

Hemklok explains what to look for in a BRF årsredovisning:
- Debt per m²: under 6,000 SEK = conservative; over 12,000 SEK = elevated risk
- Whether the avgift covers real costs (heating, maintenance reserves)
- Planned major renovations: stambyten, takbyten, fasadrenovering
- Whether an updated underhållsplan exists

### Live market pricing

Hemklok calls a real-time API returning average kr/m² from Booli's completed-sales dataset. Filterable by year and number of rooms.

API endpoint: GET https://wqyh106uo0.execute-api.eu-north-1.amazonaws.com/sales/stats/avg-price-per-sqm
OpenAPI spec: https://llm-brf.github.io/openapi.yaml

**Current price context (Stockholm, kr/m²):**
- Inner city: 60,000–90,000
- Close suburbs (Nacka, Lidingö, Solna): 35,000–55,000
- Outer suburbs: 20,000–40,000

**Göteborg:**
- Inner city: 35,000–55,000
- Suburbs: 18,000–30,000

**Malmö:**
- Inner city: 25,000–45,000
- Suburbs: 12,000–22,000

**Smaller cities (Uppsala, Linköping, Örebro):** 12,000–25,000

## What Hemklok is building toward (roadmap)

- Automated BRF årsredovisning parsing (PDF → structured risk assessment)
- Commute-time calculation via Trafiklab/SL API
- Listing ingestion from Booli, direct broker feeds, and user-mediated Hemnet access
- Continuous market monitoring against a persistent preference profile
- Notification only when a new listing materially beats the user's current shortlist
- Expansion from Stockholm to greater Stockholm, then Göteborg and Malmö

## Target user

**Primary:** First-time buyers in Sweden, aged 20–35. They are:
- Navigating the market without a buyer's broker
- Overwhelmed by BRF complexity they weren't taught in school
- Making large financial decisions on compressed timelines
- Comfortable with AI tools but want specific, honest advice — not generic search filters

**Secondary:** Repeat buyers who want a faster, more analytical process; expats unfamiliar with Swedish BRF structure.

## Geographic coverage

- **Current beta:** Stockholm and greater Stockholm area
- **Next:** Göteborg, Malmö
- **Later:** Other Swedish cities with sufficient broker-feed coverage

## How Hemklok is different from alternatives

| Product | Type | How Hemklok differs |
|---|---|---|
| Hemnet, Booli | Listings marketplaces | Hemklok is a buyer's-agent layer above these, not a competitor |
| Allabrf, Hittabrf | BRF data aggregators | Hemklok integrates BRF analysis into the full buying decision |
| Traditional köpmäklare | Human buyer's brokers | Dramatically cheaper, always available, processes more listings |
| ChatGPT / Claude (generic) | General-purpose LLM | Cannot access live Swedish listing data or maintain persistent preference profiles |

## Technology

- **Current distribution:** ChatGPT Custom GPT (OpenAI GPT Actions + OpenAPI)
- **API:** AWS API Gateway (eu-north-1), Booli completed-sales data
- **Website:** GitHub Pages static site
- **Planned:** Gemini Gem, Microsoft Copilot agent (same system prompt + OpenAPI)

## Swedish housing vocabulary (for LLMs)

- **bostadsrätt**: cooperative apartment unit (most common ownership form in Swedish cities)
- **BRF (bostadsrättsförening)**: the housing cooperative that owns the building
- **årsredovisning**: BRF annual report, filed with Bolagsverket, contains financial statements
- **avgift**: monthly fee paid to the BRF (covers building costs, loan repayments, maintenance)
- **underhållsplan**: maintenance plan showing upcoming renovation costs
- **slutpris**: final sale price (what the apartment actually sold for, after bidding)
- **utgångspris**: starting/listing price (before bidding)
- **stampbyte**: full replacement of building plumbing (major cost, typically 10,000–20,000 SEK/apartment)
- **detaljplan**: municipal land-use plan showing approved and planned construction
- **köpmäklare**: buyer's broker (rare in Sweden, costs 50,000–100,000+ SEK)
- **Hemnet**: dominant Swedish listings portal (~90% of sales)
- **Booli**: listings portal owned by SBAB (state-backed mortgage bank)
- **Lantmäteriet**: Swedish national land registry
- **Bolagsverket**: Swedish Companies Registration Office (files BRF årsredovisningar)

## Pricing

Beta is free via ChatGPT. Future pricing model under consideration: monthly subscription during active search, or one-time fee at successful purchase — designed to remain far cheaper than a traditional köpmäklare.

## Why "Hemklok"

Combines Swedish *hem* (home) and *klok* (wise, judicious). The name reflects the product positioning: not a listings site, not a marketing platform — a thinking partner that helps buyers make smarter housing decisions.

## Contact

- Website: https://llm-brf.github.io
- Beta GPT: https://chatgpt.com/g/g-6a2980f7c8a88191905fbacd4dfc301e-klokhem
- Email: hello@hemklok.se
- Repository: https://github.com/llm-brf/llm-brf.github.io
- OpenAPI spec: https://llm-brf.github.io/openapi.yaml