Platform
Get started

Casafari is the AI agent-native real estate data intelligence platform. The most complete property index in Europe: a deduplicated, cleaned property graph of residential and commercial property, for sale and for rent, in 16 countries.The most complete property index in Europe: a deduplicated, cleaned property graph.

How the graph is built

Agency partnerships in Germany · Use case

How do I get notified when a watched agency or agent changes?

Save agencies and agents from an AgentGraph AI screen to your organisation's shared watchlist with agentgraph_save_to_watchlist. AgentGraph AI re-measures the list every day and records changes such as momentum turning, more price cuts, slower exits, falling stock or an agent moving agency. Ask agentgraph_watchlist_changes on the Casafari MCP server what changed since a date.

Products: AgentGraph AI

Short answer

Your organisation has one shared watchlist in AgentGraph AI. AgentGraph AI re-measures it every day and records changes: priority or profile changes, momentum turning against the area, more price cuts, slower exits, stock falling or growing, an agent advertising under a new agency, and no newly advertised homes for 60 days. You read the changes with agentgraph_watchlist_changes; the tools describe no push delivery, so an assistant or agent asks on a schedule.

The moment a target agency loses momentum or a star agent moves is the moment to call, and the watchlist is what catches it.

Steps

Save ids from a screen: agentgraph_save_to_watchlist with location, operation, label (up to 50 ids per call, 200 entries)

  1. Call agentgraph_save_to_watchlist with ids from agentgraph_screen_agencies, agentgraph_screen_agents or a profile (1 to 50 per call), the location (or zipcodes) to measure each entry in, operation (sale by default, or rent) and an optional label up to 60 characters, such as 'Q4 outreach'. The watchlist holds up to 200 entries.

Review: agentgraph_watchlist

  1. Call agentgraph_watchlist (no inputs). It returns your organisation's watched agencies and agents with their area, latest priority, active homes and momentum against the area.

Ask what changed: agentgraph_watchlist_changes with since, min_importance, refresh

  1. Call agentgraph_watchlist_changes with:
    • since: a date as YYYY-MM-DD (default: the last 7 days).
    • min_importance: high, medium or low (default low, meaning everything).
    • refresh: true re-measures now instead of waiting for the daily run (slower); default false.

Prune: agentgraph_remove_from_watchlist

  1. Call agentgraph_remove_from_watchlist with the ids to remove. This removes them for your whole organisation.

Example: a weekly check

Prompt to your AI assistant, or a scheduled agent:

What changed on my watchlist this week? Only the important things.

The assistant calls agentgraph_watchlist_changes with arguments like these:

json
{
  "since": "(date 7 days ago, YYYY-MM-DD)",
  "min_importance": "high"
}

The answer lists changes most important first within each day. A sample layout (placeholders shown; the tool returns the real entries):

DayEntryChangeImportance
(date)(agency)Momentum turning against the area(level)
(date)(agent)Advertising under a new agency(level)
(date)(agency)No newly advertised homes for 60 days(level)

For any entry that matters, the assistant can open agentgraph_agency_profile or agentgraph_agent_profile, or prepare agentgraph_partner_brief.

How to read the result

Importance levels high, medium, low

Each change carries an importance of high, medium or low, and changes are sorted most important first within each day. Set min_importance to high for a short weekly digest, or leave it at low to see every change. A priority or profile change on a target is usually the trigger to call; a single price cut is context.

refresh true re-measures now and is slower

Without refresh, you read the changes recorded by the daily run. With refresh: true, AgentGraph AI re-measures the watchlist now, which takes longer. Use it before an important call, not in a routine weekly check.

Limits

AgentGraph AI tools are available to Casafari clients with MCP access, using their Casafari account; self-serve access for new users is coming soon. They are MCP only, with no REST equivalent. The sources give no prices, quotas, rate limits or SLAs; ask Casafari.

Watchlist is shared across your organisation

There is one watchlist per organisation. Everyone in it sees the same entries, and a removal applies to all.

Germany only

AgentGraph AI covers Germany: each entry is measured in a German area or set of five-digit postcodes.

Up to 200 entries

The watchlist holds up to 200 entries, saved at most 50 ids per call. Remove entries you no longer need to make room.

Questions

Can I get watchlist changes pushed to me?

The AgentGraph AI tools describe no push, email or webhook delivery for watchlist changes, so you ask for them. Have an assistant or a scheduled agent call agentgraph_watchlist_changes, for example weekly with since set to seven days ago and min_importance high. The webhooks on the hub belong to New Market Leads alert feeds over REST, not to the AgentGraph AI watchlist.

What counts as going quiet?

Going quiet means a watched agency or agent has had no newly advertised homes for 60 days, as the agentgraph_save_to_watchlist description states. AgentGraph AI records it as a change on the watchlist, alongside other changes such as momentum turning, more price cuts, slower exits and falling stock. Read it with agentgraph_watchlist_changes.

Can I watch the same agency in two areas?

Each watchlist entry is measured in the area you give when you save it with agentgraph_save_to_watchlist. The tool description does not say whether one id can be saved twice with two different areas, so this page cannot confirm it. Check the result with agentgraph_watchlist after saving, or ask Casafari at https://www.casafari.com/.

Who in my organisation sees the watchlist?

The watchlist is your organisation's shared watchlist, so everyone in the organisation using AgentGraph AI sees the same entries and the same changes. agentgraph_watchlist lists them with area, latest priority, active homes and momentum. Removing an entry with agentgraph_remove_from_watchlist removes it for everyone, so agree who prunes the list.

All use cases

Tip: add .md to any URL to read it as Markdown.