Docs/Custom Actions/Custom Actions

    Examples

    Last updated · MAR 2026·Read as Markdown

    Integration Examples

    Common Custom Action patterns and real-world AV integrations.

    Note: Custom Actions require a Builder plan or higher. Upgrade to Builder

    Push status updates from the agent to a Slack channel via webhook. Useful for install-team standups, commissioning callouts, or escalations.

    Configuration#

    bash
    curl -X POST https://hooks.slack.com/services/{{var.SLACK_WEBHOOK_ID}} \
  -H "Content-Type: application/json" \
  -d '{
    "text": "{{message}}",
    "channel": "{{channel}}",
    "username": "AVCodex Bot"
  }'

    Parameters#

    Parameter Source Configuration
    message AI Generated Message content to send.
    channel AI Generated Target channel (optional).
    SLACK_WEBHOOK_ID Variable Your Slack webhook ID.

    Usage#

    _"Post to #av-ops that Boardroom 4 commissioning passed."_

    Call OpenAI's API for an extra round of analysis (for example, summarizing a long commissioning checklist).

    Configuration#

    bash
    curl -X POST https://api.openai.com/v1/chat/completions \
  -H "Authorization: Bearer {{var.OPENAI_API_KEY}}" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4",
    "messages": {{system.message_history}},
    "temperature": 0.7,
    "max_tokens": {{maxTokens}}
  }'

    Parameters#

    Parameter Source Configuration
    OPENAI_API_KEY Variable (secret) Your OpenAI API key.
    message_history System variable Conversation context.
    maxTokens AI Generated Token limit (default: 1000).

    Send a transactional email, for example a client-acceptance email after commissioning.

    Configuration#

    bash
    curl -X POST https://api.sendgrid.com/v3/mail/send \
  -H "Authorization: Bearer {{var.SENDGRID_API_KEY}}" \
  -H "Content-Type: application/json" \
  -d '{
    "personalizations": [{
      "to": [{"email": "{{recipientEmail}}"}],
      "subject": "{{subject}}"
    }],
    "from": {
      "email": "{{var.FROM_EMAIL}}",
      "name": "{{var.FROM_NAME}}"
    },
    "content": [{
      "type": "text/html",
      "value": "{{emailBody}}"
    }]
  }'

    Variables#

    bash
    SENDGRID_API_KEY = SG.abc123...
FROM_EMAIL = noreply@your-integrator.com
FROM_NAME = Your Integrator

    Read or append to a Google Sheet. Common use: an asset register or a per-room commissioning log.

    Read data#

    bash
    curl -X GET "https://sheets.googleapis.com/v4/spreadsheets/{{var.SHEET_ID}}/values/{{range}}" \
  -H "Authorization: Bearer {{var.GOOGLE_API_KEY}}"

    Append data#

    bash
    curl -X POST "https://sheets.googleapis.com/v4/spreadsheets/{{var.SHEET_ID}}/values/{{range}}:append" \
  -H "Authorization: Bearer {{var.GOOGLE_API_KEY}}" \
  -H "Content-Type: application/json" \
  -d '{
    "values": [[
      "{{system.timestamp}}",
      "{{system.user_id}}",
      "{{data}}"
    ]]
  }'

    Multi-step workflow using action dependencies. Example: register a new site, then push a welcome ticket to the field-service queue.

    Step 1: create site record#

    bash
    curl -X POST https://api.example.com/sites \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer {{var.API_KEY}}" \
  -d '{
    "name": "{{siteName}}",
    "client": "{{clientName}}"
  }'

    Returns: {"data": {"id": "site_123", "name": "1200 Market"}}.

    Step 2: open kickoff ticket#

    Use the site ID from Step 1:

    bash
    curl -X POST https://api.example.com/tickets \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer {{var.API_KEY}}" \
  -d '{
    "site_id": "{{siteId}}",
    "template": "kickoff",
    "send_at": "{{system.timestamp}}"
  }'

    Parameter configuration for `site_id`:

    • Source: Output from another Action.
    • Action: Create site record.
    • JSONPath: $.data.id.

    Update client records in your CRM (HubSpot, Salesforce, ConnectWise).

    Configuration#

    bash
    curl -X PATCH https://api.crm.com/v2/contacts/{{contactId}} \
  -H "Authorization: Bearer {{var.CRM_API_KEY}}" \
  -H "Content-Type: application/json" \
  -d '{
    "properties": {
      "last_contact": "{{system.timestamp}}",
      "last_message": "{{summary}}",
      "status": "{{status}}"
    }
  }'

    AI instructions#

    For summary:

    code
    Summarize the key points of the conversation in 1-2 sentences. Mention any rooms, devices, or deadlines referenced.

    For status:

    code
    Choose one: 'rfp-open', 'design-review', 'install-active', 'commissioning', 'closed-won', 'closed-lost'.

    Create payment intents via Stripe, for example accepting a deposit on a hospitality AV refresh.

    Configuration#

    bash
    curl -X POST https://api.stripe.com/v1/payment_intents \
  -u "{{var.STRIPE_SECRET_KEY}}:" \
  -H "Content-Type: application/x-www-form-urlencoded" \
  -d "amount={{amount}}&currency=usd&customer={{customerId}}"

    Parameters#

    Parameter Source Configuration
    STRIPE_SECRET_KEY Variable (secret) Your Stripe secret key.
    amount AI Generated Amount in cents.
    customerId AI Generated Stripe customer ID.

    Pull a device's recent status from a monitoring tool, for example checking a Q-SYS Core or a Crestron 4-Series before dispatching a tech.

    Configuration#

    bash
    curl -X GET "https://api.your-monitoring.com/devices/{{deviceId}}/status" \
  -H "Authorization: Bearer {{var.MONITORING_API_KEY}}"

    Usage examples#

    _"Is the Q-SYS Core in the auditorium online?"_ _"What was the last reboot time for the boardroom Crestron processor?"_

    Open issues in a repo, for example a control-system code library where programmers track Crestron SIMPL or Q-SYS Lua modules.

    Configuration#

    bash
    curl -X POST https://api.github.com/repos/{{var.GITHUB_OWNER}}/{{var.GITHUB_REPO}}/issues \
  -H "Authorization: Bearer {{var.GITHUB_TOKEN}}" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "{{issueTitle}}",
    "body": "{{issueBody}}",
    "labels": {{labels}},
    "assignees": {{assignees}}
  }'

    Parameter configuration#

    Parameter Source AI instructions
    issueTitle AI Generated Concise issue title that names the module and symptom.
    issueBody AI Generated Detailed description, include processor model and firmware.
    labels AI Generated Array of labels, for example ["crestron", "simpl", "bug"].
    assignees AI Generated Array of GitHub usernames to assign.

    1. Error handling#

    Always return clear error messages:

    json
    {
  "success": false,
  "error": "API key invalid. Please check your credentials."
}

    2. Rate limiting#

    Respect API rate limits:

    • Add delays between requests.
    • Implement exponential backoff.
    • Cache responses when possible.

    3. Security#

    • Use HTTPS endpoints only.
    • Store credentials as secret variables.
    • Never log sensitive data.
    • Validate all inputs.

    4. Performance#

    • Paginate large result sets.
    • Use field selection to minimize data.
    • Implement timeouts (30s max).
    • Cache frequently accessed data.

    5. Testing#

    • Test with various inputs.
    • Verify error handling.
    • Check edge cases.
    • Monitor response times.

    *AVCodex · Your AV expertise. Amplified by AI.*

    Was this helpful?
    Edit this page →