AdCP 3.1.13 · Signals
get_signals
get_signals is an operation in the Signals area of AdCP 3.1.13. get_signals is the AdCP task for discovering audience and contextual signals.
get_signals request fields
The required column reflects the top-level required array only. The request schema declares no required array, so none of the 18 request fields carries a top-level required flag. A top-level if, anyOf, oneOf or dependencies block can make an unmarked field mandatory in the mode being called, and a flat table cannot render
a rule that fires in one mode only. Those branches are visible in the request schema itself.
| Field | Type | Required | Description |
|---|---|---|---|
adcp_version | string | Release-precision AdCP version (VERSION.RELEASE, e.g. | |
adcp_major_version | integer | DEPRECATED in favor of adcp_version (release-precision string). | |
discovery_mode | string | Declares caller intent for this request. One of: brief, wholesale. | |
account | account-ref | Account for this request. | |
signal_spec | string | Natural language description of the desired signals. | |
signal_refs | signal-ref[] | Specific signals to look up by reference. | |
signal_ids | signal-id[] | DEPRECATED. | |
destinations | destination[] | Filter signals to those activatable on specific agents/platforms. | |
countries | string[] | Countries where signals will be used (ISO 3166-1 alpha-2 codes). | |
filters | signal-filters | Filters to refine signal discovery results | |
fields | string[] | Specific signal fields to include in the response, aligned with get_products.fields. One of: signal_ref, signal_id, signal_agent_segment_id, name, description, value_type, categories, range, signal_type, data_provider, coverage_percentage, deployments, pricing_options, taxonomy, data_sources, methodology, segmentation_criteria, criteria_url, refresh_cadence, lookback_window, onboarder, modeling, audience_expansion, device_expansion. | |
max_results | integer | DEPRECATED: Use pagination.max_results instead. | |
pagination | pagination-request | Pagination parameters. | |
push_notification_config | push-notification-config | Optional webhook configuration for async terminal completion/failure notifications on semantic signal discovery. | |
if_wholesale_feed_version | string | Opaque wholesale_feed_version token returned by a prior wholesale-mode get_signals response from this agent. | |
if_pricing_version | string | Opaque pricing_version token from a prior get_signals response. | |
context | context | Opaque correlation data that is echoed unchanged in responses. | |
ext | ext | Extension object for platform-specific, vendor-namespaced parameters. |
get_signals reads. 30 of the 64 registered operations are flagged x-mutates-state and its request schema is not one of them, so repeating the call is side-effect free. AdCP's release notes describe x-mutates-state as a non-normative tooling
hint that agents do not validate, so it is the registry's classification of the operation rather than
a contract, and it is the only machine-readable answer in the registry to whether an operation is
safe to retry.
get_signals response fields
The first 13 rows are the shared protocol envelope from core/version-envelope.json and core/protocol-envelope.json. Only 10 of the 23 rows below belong to get_signals itself, beginning with signals, errors and incomplete. One field in the whole response is marked required: status.
| Field | Type | Required | Description |
|---|---|---|---|
adcp_version | string | Release-precision AdCP version (VERSION.RELEASE, e.g. | |
adcp_major_version | integer | DEPRECATED in favor of adcp_version (release-precision string). | |
context_id | string | Session/conversation identifier for tracking related operations across multiple task invocations. | |
context | context | Opaque correlation data that is echoed unchanged in responses. | |
task_id | string | Unique identifier for tracking asynchronous operations. | |
status | task-status | required | Current task execution state. One of: submitted, working, input-required, completed, canceled, failed, rejected, auth-required, unknown. |
message | string | Human-readable summary of the task result. | |
timestamp | string | ISO 8601 timestamp when the response was generated. | |
replayed | boolean | Set to true when this response was returned from the idempotency cache rather than from a fresh execution. | |
adcp_error | error | Transport-envelope error signal for fatal task failures. | |
push_notification_config | push-notification-config | Push notification configuration for async task updates (A2A and REST protocols). | |
governance_context | string | Governance context token issued by the account's governance agent during check_governance. | |
payload | object | Conceptual grouping for the task-specific response data defined by individual task response schemas (e.g., get-products-response.json, create-media-buy-response.json). | |
signals | object[] | Array of matching signals | |
errors | error[] | Task-specific errors and warnings (e.g., signal discovery or pricing issues) | |
incomplete | object[] | Declares what the agent could not finish within the caller's time_budget or due to internal limits. | |
wholesale_feed_version | string | Opaque token representing the version of the wholesale signals feed state used to compose this response. | |
pricing_version | string | Opaque token representing the version of the pricing layer. | |
cache_scope | string | Declares whether the wholesale_feed_version and pricing_version on this response describe a universal layer or an account-specific overlay. One of: public, account. | |
unchanged | boolean | Present and `true` ONLY on wholesale-mode responses when the request carried if_wholesale_feed_version (and/or if_pricing_version) matching the agent's current version for the caller's cache_scope, in which case signals… | |
pagination | pagination-response | Standard cursor-based pagination metadata for list responses | |
sandbox | boolean | When true, this response contains simulated data from sandbox mode. | |
ext | ext | Extension object for platform-specific, vendor-namespaced parameters. |
The shared protocol envelope
62 of AdCP's 64 registered operations carry every one of the 13 envelope fields. report_plan_outcome and check_governance compose core/version-envelope.json only, so they carry the version pair and none of the task fields.
status resolves to enums/task-status.json, which admits nine values:
submitted, working, input-required, completed, canceled, failed, rejected, auth-required,
unknown. The envelope marks it required on every response, and a synchronous call emits completed rather than omitting it. When it returns submitted or working instead, task_id is the
polling handle and
push_notification_config is the webhook the agent echoes back to confirm the callback.
Webhooks are A2A and REST only: MCP sends progress notifications and defines no webhook.
payload appears in the response field table typed object. core/protocol-envelope.json describes it as a documentary construct and states that it is not a required wire field. Body fields
are siblings of the envelope fields: on MCP at the root of the tool response, on A2A in task.artifacts[0].parts[].DataPart, on REST at the root of the JSON body. An accessor that reads response.payload finds
nothing there.
Field lists are not published over the wire
AdCP's integration guide, docs/protocol/calling-an-agent.mdx, states that its MCP
servers no longer publish per-tool parameter schemas: a tools/list call returns an empty
properties object for every tool, and the guide directs clients not to infer shape from it. A client
author reads the schema registry instead of the tool list.
Example get_signals calls
The field tables above come from the 3.1.13 schema release and these blocks come from the documentation tree, which is versioned separately, so the two can disagree.
{
"signals": [
{
"signal_ref": {
"scope": "data_provider",
"data_provider_domain": "string",
"signal_id": "string"
},
"signal_agent_segment_id": "string",
"name": "string",
"description": "string",
"signal_type": "string",
"data_provider": "string",
"coverage_percentage": "number (optional, deprecated)",
"coverage_forecast": {
"method": "estimate",
"forecast_range_unit": "availability",
"scope": {
"kind": "inventory",
"label": "network price-priority inventory"
},
"bucket_semantics": "exclusive",
"bucket_completeness": "partial",
"points": [
{
"label": "not present",
"dimensions": [
{
"kind": "signal",
"signal_ref": {
"scope": "data_provider",
"data_provider_domain": "weather-data.example",
"signal_id": "weather"
},
"signal_value": null,
"presence": "absent"
}
],
"metrics": {
"impressions": { "mid": 280000 },
"coverage_rate": { "mid": 0.28 }
}
},
{
"label": "hot",
"dimensions": [
{
"kind": "signal",
"signal_ref": {
"scope": "data_provider",
"data_provider_domain": "weather-data.example",
"signal_id": "weather"
},
"signal_value": "hot",
"presence": "present"
}
],
"metrics": {
"impressions": { "mid": 180000 },
"coverage_rate": { "mid": 0.18 }
}
}
]
},
"deployments": [
{
"type": "agent",
"agent_url": "string",
"account": "string",
"is_live": "boolean",
"activation_key": {
"type": "segment_id",
"segment_id": "string"
},
"estimated_activation_duration_minutes": "number"
}
],
"pricing_options": [
{
"pricing_option_id": "string",
"model": "cpm | percent_of_media | flat_fee | per_unit | custom",
"...": "..."
}
]
}
]
} {
"type": "segment_id",
"segment_id": "ttd_segment_12345"
} {
"type": "key_value",
"key": "audience_segment",
"value": "luxury_auto_intenders"
} Frequently asked
- What does get_signals do?
- get_signals is the AdCP task for discovering audience and contextual signals.
- Which fields does a get_signals request require?
- The published 3.1.13 schema declares no required array on the get_signals request. Constraints can still apply: a top-level if, anyOf, oneOf or dependencies rule fires in one mode only, and a flat field table cannot render it.
- Which AdCP version do these get_signals field tables describe?
- AdCP 3.1.13, the current published release, last updated 2026-08-11. The documentation tree is versioned separately from the schema tree in the same repository, so a doc page and a schema file can describe the same field differently.