> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vortexiq.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Google Gemini on Vortex IQ

> Monitor Google Gemini health, cost and reliability signals, and catch incidents and runaway spend early.

Monitor Google Gemini health, cost and reliability signals, and catch incidents and runaway spend early.

[Connect or manage this source](https://app.vortexiq.ai/workbench/settings/sources) · [How connecting works](/integrations/connector-catalogue) · [Create a workflow](https://app.vortexiq.ai/workbench/flows/create?connector=gemini)

<CardGroup cols={5}>
  <Card title="14">
    performance signals
  </Card>

  <Card title="3">
    automated checks
  </Card>

  <Card title="Build your own">
    automated fixes
  </Card>

  <Card title="17">
    proven workflows
  </Card>

  <Card title="1">
    API operations
  </Card>
</CardGroup>

<Tabs>
  <Tab title="Overview">
    ### What you can achieve

    Capabilities are grouped around merchant outcomes, not API terminology.

    <CardGroup cols={2}>
      <Card title="Catalogue quality">
        Identify missing, inconsistent or outdated product data.
      </Card>

      <Card title="Control risk and change">
        Keep tracking, access and change under governed control.
      </Card>

      <Card title="Customer experience">
        Find storefront, speed, accessibility and journey problems.
      </Card>

      <Card title="Run operations">
        Monitor orders, fulfilment, delivery and settlement.
      </Card>
    </CardGroup>

    ### From connection to verified outcome

    The controlled sequence every capability follows. Nothing changes a connected system without the approval step.

    <Steps>
      <Step title="Connect">
        Authorise the source. Scopes are shown before access is granted.
      </Step>

      <Step title="Monitor">
        Watch the signals against your own baselines, not universal defaults.
      </Step>

      <Step title="Detect">
        Run checks and gather evidence specific to your store.
      </Step>

      <Step title="Recommend">
        Explain what happened, why it matters and the proposed action.
      </Step>

      <Step title="Approve">
        You review scope, risk and reversibility before anything changes.
      </Step>

      <Step title="Execute">
        Apply through governed connector operations.
      </Step>

      <Step title="Verify">
        Confirm the intended result and keep the receipt.
      </Step>
    </Steps>

    No changes are made without the configured approval policy. Read-only operations do not modify the connected system; schedules, access scopes, API usage and data handling remain governed by Vortex IQ controls.
  </Tab>

  <Tab title="Monitor (14)">
    ### Monitor performance

    14 performance signals. Open an outcome to see its signals and how each one alerts. Read-only operations do not modify the connected system.

    <AccordionGroup>
      <Accordion title="Catalogue quality (8 signals)">
        | Signal                            | Alert behaviour    | What it tracks                                                                                                                                         |
        | --------------------------------- | ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------ |
        | **Avg Input Token Limit**         | Watch only         | Mean of inputTokenLimit over models that report one, from /v1beta/models.                                                                              |
        | **Gemini API Health**             | Alert band 90 / 70 | Composite 0-100: 60 for the API being reachable and the key authenticating on /v1beta/models, plus up to 40 for a non-empty model catalogue. No spend  |
        | **Latest Model**                  | Watch only         | displayName (or name) of the model with the highest numeric leading version parsed from the version field across /v1beta/models.                       |
        | **Max Context Window**            | Watch only         | Maximum inputTokenLimit across all models in /v1beta/models (largest input context available).                                                         |
        | **Max Output Tokens**             | Watch only         | Maximum outputTokenLimit across all models in /v1beta/models.                                                                                          |
        | **Model Families**                | Watch only         | Distinct model families from /v1beta/models, grouping name by its family stem (e.g. gemini-2.5, gemini-2.0, gemini-1.5, text-embedding). Counted per f |
        | **Preview / Experimental Models** | Watch only         | Count of models whose name or version contains 'preview' or 'exp' from /v1beta/models (not yet GA).                                                    |
        | **Vision-Capable Models**         | Watch only         | Count of models whose description or displayName matches a multimodal keyword (vision, image, audio, multimodal) from /v1beta/models. Heuristic: Gemin |
      </Accordion>

      <Accordion title="Control risk and change (3 signals)">
        | Signal                       | Alert behaviour    | What it tracks                                                                                                                                           |
        | ---------------------------- | ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | **Gemini Spend (30d)**       | Alert band 10 / 50 | Gemini spend is NOT exposed by generativelanguage.googleapis.com; it is billed through Google Cloud. Degrades with reason billing\_via\_gcp unless a lin |
        | **Gemini Usage Trend (30d)** | Alert band 0 / -50 | Gemini request/token usage is NOT exposed by the Gemini API (no usage endpoint). Degrades with reason billing\_via\_gcp unless a linked GCP billing expo |
        | **Models Available**         | Watch only         | Count of entries in GET /v1beta/models the key can see. Renders with any valid Gemini API key.                                                           |
      </Accordion>

      <Accordion title="Customer experience (3 signals)">
        | Signal                     | Alert behaviour | What it tracks                                                                                  |
        | -------------------------- | --------------- | ----------------------------------------------------------------------------------------------- |
        | **countTokens Models**     | Watch only      | Count of models whose supportedGenerationMethods includes countTokens, from /v1beta/models.     |
        | **embedContent Models**    | Watch only      | Count of models whose supportedGenerationMethods includes embedContent, from /v1beta/models.    |
        | **generateContent Models** | Watch only      | Count of models whose supportedGenerationMethods includes generateContent, from /v1beta/models. |
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Audit (3)">
    ### Audit risks and opportunities

    A fix status appears only where the action, inputs, approval, verification and recovery controls are mapped. Candidate remediations are never executable. Open a check for the detail.

    <AccordionGroup>
      <Accordion title="API key invalid, expired or lacking the scopes the cards need">
        **Severity** high · **Outcome** Control risk and change · **Fix status** Report only

        Once this credential fails, VortexIQ stops seeing usage, cost and health data from this platform entirely, and every finding here that depends on it goes stale from that moment, silently, until someone notices the numbers stopped updating.

        Vortex IQ detects and explains this; resolution is manual, with evidence and recommended steps.

        Reference: `CLOUD-AUTH-001`
      </Accordion>

      <Accordion title="Spend up more than 50% vs the prior period">
        **Severity** high · **Outcome** Run operations · **Fix status** Report only

        Spend on this platform rose more than 50% versus the prior period. A jump this size is either a deliberate scale-up worth confirming was intentional, or an unbounded process, a stuck job, a misconfigured autoscaler, quietly running up a bill nobody approved.

        Vortex IQ detects and explains this; resolution is manual, with evidence and recommended steps.

        Reference: `CLOUD-SPEND-001`
      </Accordion>

      <Accordion title="Usage dropped more than 50% vs the prior period (silent stall)">
        **Severity** medium · **Outcome** Control risk and change · **Fix status** Report only

        Usage halving against the prior period is the earliest sign of a silent stall: a broken schedule, an expired credential, or a team that quietly stopped.

        Vortex IQ detects and explains this; resolution is manual, with evidence and recommended steps.

        Reference: `CLOUD-USAGE-001`
      </Accordion>
    </AccordionGroup>

    #### Build your own automated fixes

    Turn any finding into an automated fix with a Vortex IQ workflow: **over 13,000 read and write operations across more than 200 connectors** are available as building blocks, with approval, verification and rollback on every change.
  </Tab>

  <Tab title="Automate">
    ### Automate approved work

    Vortex IQ is integrated with **1 read** and **0 write** operations across v1betamodels on Google Gemini. Combine them with anything from the **over 13,000 operations across more than 200 connectors** to automate the work in your own words.

    Changes follow your configured approval policy: the target, proposed change, affected records, risk, reversibility and verification plan are shown before execution.

    [Create a workflow](https://app.vortexiq.ai/workbench/flows/create?connector=gemini)

    #### Proven workflow templates (17)

    <AccordionGroup>
      <Accordion title="Brand Voice Compliance Review">
        **For** Product content manager · **Ready to deploy**

        flags off-brand, unsupported or inconsistent product claims and proposes corrections.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Catalogue Completeness Guard">
        **For** Product content manager · **Ready to deploy**

        scores required fields by category and routes the highest-revenue gaps first.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Duplicate Copy Resolver">
        **For** Product content manager · **Ready to deploy**

        detects reused descriptions and produces differentiated drafts from verified attributes.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="New SKU Content Factory">
        **For** Product content manager · **Ready to deploy**

        turns approved source data into complete draft PDPs with a human review queue.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="PDP Enrichment Queue">
        **For** Product content manager · **Ready to deploy**

        finds thin product pages, generates on-brand copy, and batches changes for approval.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Product Claim Evidence Check">
        **For** Product content manager · **Ready to deploy**

        identifies unsubstantiated claims and requires evidence or removal before publish.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Seasonal Content Refresh">
        **For** Product content manager · **Ready to deploy**

        identifies stale seasonal language and prepares date-safe updates with expiry/rollback.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Supplier Copy Normaliser">
        **For** Product content manager · **Ready to deploy**

        converts inconsistent feeds into house style while preserving factual specifications.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Variant Consistency Sweep">
        **For** Product content manager · **Ready to deploy**

        reconciles names, options, specifications and copy across product families.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Brief Approval Router">
        **For** Content manager · **Ready to deploy**

        sends briefs to the correct legal, SEO, product and brand approvers based on content type.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, HubSpot, Mailchimp, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Brief-to-Draft Pipeline">
        **For** Content manager · **Ready to deploy**

        produces evidence-backed article drafts with sources, brand rules and approval status.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, HubSpot, Mailchimp, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Campaign Asset Repurposer">
        **For** Content manager · **Ready to deploy**

        turns an approved core message into channel-specific drafts without auto-publishing.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, HubSpot, Mailchimp, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Content Calendar Builder">
        **For** Content manager · **Ready to deploy**

        creates a channel-ready editorial calendar from launches, seasonality and search demand.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, HubSpot, Mailchimp, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Content Freshness Watch">
        **For** Content manager · **Ready to deploy**

        finds decaying or outdated pages, explains why they matter and opens refresh tasks.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, HubSpot, Mailchimp, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Editorial QA Gate">
        **For** Content manager · **Ready to deploy**

        checks links, claims, tone, formatting, metadata and calls to action before approval.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, HubSpot, Mailchimp, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Evergreen Refresh Queue">
        **For** Content manager · **Ready to deploy**

        schedules evidence and date checks for high-value guides.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, HubSpot, Mailchimp, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>

      <Accordion title="Launch Story Pack">
        **For** Content manager · **Ready to deploy**

        creates coordinated blog, email and social drafts from verified product information.

        Also available on: OpenAI, DigitalOcean, Google Cloud Platform, Amazon Web Services, Microsoft Azure, Anthropic (Claude)

        Tested alongside: BigCommerce, Shopify, HubSpot, Mailchimp, Anthropic

        Runs on a schedule or trigger. Any step that changes data pauses for approval, and results deliver to the configured destination with a run receipt.
      </Accordion>
    </AccordionGroup>

    <Accordion title="Browse the operations you can build with">
      | Resource     | Read operations | Write operations |
      | ------------ | --------------- | ---------------- |
      | v1betamodels | 1               | 0                |

      Signed-in users see the full catalogue in the workflow builder, filtered to the sources they have connected.
    </Accordion>

    3 further designed workflows are in validation and become activatable only once proven against a live account. Prototypes are never presented as ready.
  </Tab>
</Tabs>