klados

KG Cluster

01KJ60VSSCSQQ32PRJS0BQJ6Q8

Properties

accepts

actions_required

description

Clusters semantically similar entities into hierarchical groups using relationship-based discovery and size-bounded clustering.

endpoint

https://kg-cluster.nick-chimicles-professional.workers.dev

endpoint_verified_at

2026-02-23T19:50:58.337Z

input_schema

description

Configurable parameters for clustering behavior. All have sensible defaults. Properties map directly to keys in request.input.

properties

follower_poll_interval_ms

follower_wait_max_ms

default: 60000
description: Maximum time to wait for followers after creating a cluster (ms).
minimum: 0
required: false
type: integer

follower_wait_min_ms

default: 30000
description: Minimum time to wait for followers after creating a cluster (ms). Uses jittery wait between min and max.
minimum: 0
required: false
type: integer

index_wait_ms

default: 90000
description: Minimum wait time for semantic indexing before searching (ms). Entities must be indexed to be found via similarity search.
minimum: 0
required: false
type: integer

initial_spread_ms

default: 30000
description: Additional random spread for staggering concurrent job starts (ms). Total initial delay = index_wait_ms + random(0, initial_spread_ms).
minimum: 0
required: false
type: integer

default: 15
description: Number of similar peers to consider for clustering. Higher K finds more potential clusters but increases API calls.
maximum: 50
minimum: 1
required: false
type: integer

max_cluster_size

default: 5
description: Soft cap on cluster members before overflow. Acts as branching factor - hierarchy depth is log_max(N). Smaller values create deeper, more specific hierarchies.
maximum: 50
minimum: 2
required: false
type: integer

match

produces

status

active