Resources
Full reference for every Teads MCP resource, with parameters, response shapes, and example prompts.
The Teads MCP server exposes resources (catalogs the assistant can read). This page documents every resource for people who want to understand exactly what the assistant can read.
Resources
Read these when your client supports MCP resources. If it cannot, use the matching list_* tool
instead.
| URI | Purpose |
|---|---|
teads://my-seats | Same data as list_my_seats |
teads://reporting/dimensions | Same data as list_reporting_dimensions |
teads://reporting/metrics | Same data as list_reporting_metrics |
teads://taxonomy/device-types | Device type ID catalog |
teads://taxonomy/platform-environment | Targeted platform environment values (Web, App) |
teads://taxonomy/channel-type | Inventory channel type values |
teads://taxonomy/website-tiers | Website tier values (Select, Reach, Premier) |
teads://taxonomy/ad-positions | Ad placement position values |
teads://taxonomy/target-positions | Line item target position values |
teads://taxonomy/brand-safety-providers | Brand safety provider names |
teads://taxonomy/viewability-measurement | Viewability measurement settings and primary provider |
teads://taxonomy/device-languages | Device language codes |
teads://taxonomy/operating-systems | Targetable operating system families |
teads://taxonomy/browser-types | Targetable browser types |
teads://taxonomy/connection-types | Targetable connection type values |
teads://taxonomy/website-categories | IAB website category catalog |
teads://taxonomy/format-positions | Active ad format position catalog |
teads://taxonomy/audiences/{channel} | Standard audiences; {channel} is web or ctv |
teads://taxonomy/contextual/{channel} | Contextual segment catalog; {channel} is web or ctv |
teads://taxonomy/content-languages/{channel} | Content languages catalog; {channel} is web or ctv |
teads://taxonomy/brand-suitability/{provider} | Brand suitability segments; {provider} is double-verify |
teads://knowledge/index | Same data as list_knowledge |
teads://knowledge/playbooks/{name} | Same data as read_knowledge |
Common workflows
The server ships built-in playbooks for these — the assistant follows them automatically, but they're a useful map of how the tools fit together.
Campaign lookup. list_my_seats (or get_campaign_seat if you only have the campaign ID) →
campaign_read → list_taxonomy to resolve any targeting IDs in the response.
Custom report export. Check field names with list_reporting_dimensions /
list_reporting_metrics → reporting_create (or reporting_create_task) → reporting_wait_for
for the download link. Check in-flight jobs any time with get_running_reports.
End-of-campaign report. get_campaign_seat → campaign_read to read the setup and classify
the campaign by its advertising solution → run the matching set of reporting_create calls
(core delivery, plus solution-specific and standard breakdown reports) → summarize delivery versus
what was booked. Ask for a wrap-up report on a finished campaign and the assistant runs this
automatically.
Audience insights. audience_insights to query readership intelligence by campaign entity,
custom audience, data-provider segment, or device profile filtered by category.
Reach measurement. campaign_read to get the flight dates and delivery countries →
measurement_campaign_reach_dashboard_task (or _submit → _status → _result) for unique
reach, frequency, and — with a TV reach estimate supplied — incremental reach over linear TV.
Timing and seasonality. planning_media_barometer_task (or _media_barometer → _get_status
→ _get_result) with free keywords or contextual taxonomy items to find engagement peaks.
How assistants should use these
List seats before campaign or report work. For custom reports, prefer the task tool
(reporting_create_task) when available; otherwise call reporting_create then poll
reporting_wait_for — the create result never contains row data. The same pattern applies to the
reach dashboard and media barometer: reach for the _task tool first, and fall back to
submit/poll/result only when the client doesn't support MCP Tasks.