Shared optional inputs
All endpoint tools expose endpoint parameters as top-level typed inputs. They also share the same projection controls so you can keep payloads small, save tokens, and keep results deterministic.| Parameter | Type | Description |
|---|---|---|
projection_preset | string | Optional. Defaults to minimal. Allowed values: full, minimal, identity, engagement, content. |
data_fields | string[] | Optional. Adds top-level keys from data on top of the selected projection_preset. |
item_fields | string[] | Optional. Only on collection tools. Adds keys from data.orderedItems[] on top of the selected projection_preset. |
Projection options
Use projections to keep tool results focused on the data your agent actually needs. Smaller payloads mean fewer tokens consumed per call.| Preset | Best for | What it keeps |
|---|---|---|
full | Full payload access | Returns the full API response without projection. |
minimal | Default lookups | Keeps a small set of commonly useful identity fields. On collection tools, it also keeps paging fields such as cursor, nextCursor, itemCount, and orderedItems. |
identity | Profiles and account metadata | Focuses on identity and profile fields such as names, bios, verification state, and audience counts. |
engagement | Metrics and comparisons | Focuses on counts such as views, likes, comments, shares, reposts, saves, and follower totals. |
content | Media and captions | Focuses on content fields such as text, media attachments, tags, duration, authorship, and language. |
Adding fields on top
- Use
data_fieldsto add top-level keys fromdataon top of the selected preset. - Use
item_fieldson collection tools to add keys from each item indata.orderedItems[]. - If you know exactly what you need, start with
minimaland add only the extra fields required for the task.
Response shape
Successful tool calls keep the API envelope underdata and add MCP credit
metadata when available: