> ## Documentation Index > Fetch the complete documentation index at: https://docs.adside.ai/llms.txt > Use this file to discover all available pages before exploring further. # Get Insights > LIVE performance insights straight from the Meta Graph API (impressions, clicks, spend, CTR, CPC, actions) for a campaign, ad set, ad, or account. Supports date presets, custom date ranges, breakdowns, and custom field selection. If reporting-warehouse tools are available to you in this session (`get_top_creatives_data`, `get_campaign_spend`, `get_account_summary`, `query_reporting`), this tool is the fallback, not the default. Use it only for what the warehouse cannot answer: today and the last 24h, or a breakdown it does not carry. Ranking ads or campaigns over a past period ('which ads should I kill/scale', 'top creatives'), multi-day totals, and week- or month-over-month trends all belong to the warehouse — it aggregates in SQL, answers for every object in one call, and returns a ranked handful of rows instead of one row per object. This tool returns one raw JSON row per object per time bucket, each carrying full `actions` and `cost_per_action_type` arrays. At `level: 'ad'` over a multi-day range that is tens of thousands of tokens — enough to fill your context and truncate your own answer. If you call it anyway, narrow it: the smallest `limit` that answers the question, an explicit `fields` list, and one call with `time_increment` rather than one call per period. The response is capped in size. Past that, rows are dropped and a `truncation` object says how many — the rows you get back are an arbitrary page, NOT the top performers, because this endpoint returns no ranking. `summary` is always computed over every fetched row, so totals and averages stay correct even when rows are dropped. If you see `truncation`, do not rank from what remains: ask the warehouse. ## OpenAPI ````yaml /api-reference/openapi.yaml post /meta/mcp#get_insights openapi: 3.1.0 info: title: Multi-Source Ads API version: 1.0.0 description: >- Unified API for managing ad campaigns across Meta, LinkedIn, and Google. MCP tools are accessed via JSON-RPC 2.0 over HTTP (`POST /{source}/mcp`). Google Sheets/Drive endpoints use standard REST. servers: - url: https://api.adside.ai description: Production - url: http://localhost:3100 description: Local development security: - bearerAuth: [] tags: - name: Accounts description: '' - name: Ads description: '' - name: Analytics description: '' - name: Asset Groups description: '' - name: Assets description: '' - name: Audiences description: '' - name: Campaign Groups description: '' - name: Campaigns description: '' - name: Conversions description: '' - name: Creatives description: '' - name: Drive description: '' - name: Google description: '' - name: Google Ads description: '' - name: Google Tag Manager description: '' - name: Keywords description: '' - name: Linkedin description: '' - name: Meta description: '' - name: Meta Reporting description: '' - name: OAuth description: '' - name: Other description: '' - name: Reports description: '' - name: Sheets description: '' - name: Targeting description: '' - name: Tiktok description: '' - name: Workspace description: '' paths: /meta/mcp#get_insights: post: tags: - Meta - Analytics summary: Get Insights description: >- LIVE performance insights straight from the Meta Graph API (impressions, clicks, spend, CTR, CPC, actions) for a campaign, ad set, ad, or account. Supports date presets, custom date ranges, breakdowns, and custom field selection. If reporting-warehouse tools are available to you in this session (`get_top_creatives_data`, `get_campaign_spend`, `get_account_summary`, `query_reporting`), this tool is the fallback, not the default. Use it only for what the warehouse cannot answer: today and the last 24h, or a breakdown it does not carry. Ranking ads or campaigns over a past period ('which ads should I kill/scale', 'top creatives'), multi-day totals, and week- or month-over-month trends all belong to the warehouse — it aggregates in SQL, answers for every object in one call, and returns a ranked handful of rows instead of one row per object. This tool returns one raw JSON row per object per time bucket, each carrying full `actions` and `cost_per_action_type` arrays. At `level: 'ad'` over a multi-day range that is tens of thousands of tokens — enough to fill your context and truncate your own answer. If you call it anyway, narrow it: the smallest `limit` that answers the question, an explicit `fields` list, and one call with `time_increment` rather than one call per period. The response is capped in size. Past that, rows are dropped and a `truncation` object says how many — the rows you get back are an arbitrary page, NOT the top performers, because this endpoint returns no ranking. `summary` is always computed over every fetched row, so totals and averages stay correct even when rows are dropped. If you see `truncation`, do not rank from what remains: ask the warehouse. operationId: meta_get_insights requestBody: required: true content: application/json: schema: type: object required: - jsonrpc - method - params properties: jsonrpc: type: string const: '2.0' id: type: string example: '1' method: type: string const: tools/call params: type: object required: - name - arguments properties: name: type: string const: get_insights arguments: $ref: '#/components/schemas/meta_get_insights_input' responses: '200': description: JSON-RPC response with tool result content: application/json: schema: $ref: '#/components/schemas/McpToolResponse' '401': description: Missing or invalid API key content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '403': description: Channel not connected or access denied content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '429': description: Rate limit exceeded content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' security: - bearerAuth: [] components: schemas: meta_get_insights_input: type: object properties: object_id: description: >- ID of the object to get insights for. Omit it to use the connected ad account (recommended for account-wide queries at any level). Otherwise pass a campaign ID, ad set ID, or ad ID matching `level`. Account IDs other than the connected account are rejected. Passing a page/business/user ID returns Meta error #100. type: string level: type: string enum: - account - campaign - adset - ad description: Level of insights to retrieve date_preset: description: Date preset for insights type: string enum: - today - yesterday - last_3d - last_7d - last_14d - last_28d - last_30d - last_90d - this_week_mon_today - this_week_sun_today - last_week_mon_sun - last_week_sun_sat - this_month - last_month - this_quarter - last_quarter - this_year - last_year - maximum - data_maximum time_range: description: Custom date range for insights type: object properties: since: type: string pattern: ^\d{4}-\d{2}-\d{2}$ description: Start date (YYYY-MM-DD) until: type: string pattern: ^\d{4}-\d{2}-\d{2}$ description: End date (YYYY-MM-DD) required: - since - until fields: description: Specific fields to retrieve (e.g., impressions, clicks, spend) type: array items: type: string breakdowns: description: Breakdown dimensions (e.g., age, gender, placement) type: array items: type: string time_increment: description: >- Time granularity: "monthly" for monthly rows, "all_days" for daily rows, or a number 1-90 for N-day increments anyOf: - type: string enum: - monthly - all_days - type: integer minimum: 1 maximum: 90 limit: default: 25 description: Number of insights to return type: number minimum: 1 maximum: 100 required: - level $schema: http://json-schema.org/draft-07/schema# McpToolResponse: type: object properties: jsonrpc: type: string const: '2.0' id: type: string result: type: object properties: content: type: array items: type: object properties: type: type: string const: text text: type: string description: JSON-stringified tool result ErrorResponse: type: object properties: error: type: string description: Error message code: type: string enum: - UNAUTHORIZED - FORBIDDEN - NOT_FOUND - RATE_LIMITED - INTERNAL_ERROR description: Error code securitySchemes: bearerAuth: type: http scheme: bearer description: >- API key (sk_live_*) — SHA256 hashed, looked up in Redis. Generate via the web app's workspace settings. ````