Skip to content
SupaCloud Docs
Select themeSelect language

`GET /api/intelligence/model-selection/board` — the #691 §E server-paginated model-selection board: filtered+paged rows + total + (optional) workspace facets.

GET
/api/intelligence/model-selection/board
curl --request GET \
  --url https://example.com/api/intelligence/model-selection/board

Parameters

Section titled “Parameters”

Query Parameters

Section titled “Query Parameters”
limit
integer format: int64
offset
integer format: int64
provider
string

Exact selected_agent_type (the PROVIDER facet/select).

model
string

Exact selected_model (the MODEL facet/select).

policy
string

Exact policy (the POLICY facet/select).

q
string

Free-text search over provider/model/policy + input summary.

sort
string

Whitelisted sort key: created_at | provider | model | policy (anything else folds to the default created_at DESC).

dir
string

Sort direction asc | desc (default desc).

include_facets
boolean

Recompute the workspace facets (default true; the FE sends false for a page/size/select change — the workspace facets are unchanged).

Responses

Section titled “Responses”

200

Section titled “200”
Media typeapplication/json

One page of the model-selection board: the rows, the filtered total (for the pager), and — only when requested — the workspace facets (null otherwise; the client keeps the ones it already holds).

object
facets
One of:
null
items
required
Array<object>
object
agent_profile_id
string | null format: uuid
alternatives_considered
created_at
required
string format: date-time
effort_level

ADR 0045 Stage 2b — L0 attribution + propensity (mig 212). selection_propensity is the probability the chosen arm was selected under the policy (winner_score/sum_scores for ‘auto’, 1.0 for ‘manual’, None when no arm was scored) — the inverse-propensity-weighting denominator that fixes the only-chosen-arm reward blindspot. alternatives_considered is the counterfactual log of rejected arms. effort_level/harness mirror the outcome axes for the segment join. All #[serde(default)] so pre-Stage-2 rows stay readable.

string | null
estimated_cost
required
fallback_reason
string | null
harness
string | null
id
required
string format: uuid
input_summary
required
string
override_choice
override_reason
string | null
policy
required
string
score_breakdown
required
score_completeness
required
selected_agent_type
required
string
selected_model
string | null
selection_propensity
number | null format: double
task_id
string | null format: uuid
user_overrode
required
boolean
workspace_id
required
string format: uuid
total
required
integer format: int64
Examplegenerated
{
  "facets": {
    "models": [
      {
        "count": 1,
        "value": "example"
      }
    ],
    "policies": [
      {
        "count": 1,
        "value": "example"
      }
    ],
    "providers": [
      {
        "count": 1,
        "value": "example"
      }
    ]
  },
  "items": [
    {
      "agent_profile_id": "2489E9AD-2EE2-8E00-8EC9-32D5F69181C0",
      "alternatives_considered": "example",
      "created_at": "2026-04-15T12:00:00Z",
      "effort_level": "example",
      "estimated_cost": "example",
      "fallback_reason": "example",
      "harness": "example",
      "id": "2489E9AD-2EE2-8E00-8EC9-32D5F69181C0",
      "input_summary": "example",
      "override_choice": "example",
      "override_reason": "example",
      "policy": "example",
      "score_breakdown": "example",
      "score_completeness": "example",
      "selected_agent_type": "example",
      "selected_model": "example",
      "selection_propensity": 1,
      "task_id": "2489E9AD-2EE2-8E00-8EC9-32D5F69181C0",
      "user_overrode": true,
      "workspace_id": "2489E9AD-2EE2-8E00-8EC9-32D5F69181C0"
    }
  ],
  "total": 1
}