Integrations
Integrations: what each connection reads or sends, and how to connect it today.
Broadcastwell has 22 integrations in nine groups, each connected today with a key, a webhook address, a file or a private data address you control: analytics and search, CRM, notifications and automation, reporting tools, publishing, commerce and product catalog, task management, server logs and knowledge repositories.
By the Broadcastwell team. Updated 10 November 2026. The ways to connect below are the ones every account shows on that date.
Analytics and search
| Integration | What it does | How to connect it today |
|---|---|---|
| Google Analytics | Reads the sessions your Google Analytics 4 property records from ChatGPT, Perplexity, Claude, Google AI surfaces and Copilot, by month for the last 13 months, so they sit beside your answer measurements. | Upload a GA4 export |
| Google Search Console | Reads clicks, impressions, click-through rate and average position for your top 250 search queries over the last 28 days, from the Search Console property you choose. | Upload a Queries export |
| Bing Webmaster Tools | Reads clicks and impressions for your top Bing search queries over the last 28 days, from the site you choose in Bing Webmaster Tools. | Paste your Bing Webmaster API key; Upload a search performance export |
| Adobe Analytics | Reads the visits Adobe Analytics records from AI assistant referring domains, by month for the last three months, from the report suite you choose. | Paste your Adobe server-to-server credential |
CRM
| Integration | What it does | How to connect it today |
|---|---|---|
| HubSpot | Counts the HubSpot contacts created each month whose original source was an AI assistant, from the source properties HubSpot records on each contact. Read only. | Upload a CRM export instead |
| CRM file import | Counts contacts or leads by month and AI assistant source from a CSV export of any CRM that records a created date and a source or referrer for each record. | Upload a CRM export |
Notifications and automation
| Integration | What it does | How to connect it today |
|---|---|---|
| Slack | Posts a message to a Slack channel when a run completes, an alert fires, a fix changes or a proof result changes, with a link to the receipts. | Paste a Slack incoming webhook |
| Microsoft Teams | Posts a card to a Microsoft Teams channel when a run completes, an alert fires, a fix changes or a proof result changes, through a Workflows address or an incoming webhook you create in Teams. | Paste a Teams workflow address |
| Zapier | Starts a Zap when a run completes, an alert fires or a fix changes, and lets a Zap read a run summary, preview or ask a live question, or hand off work, with an API key from your account. | Use an API key from your account |
Reporting tools
| Integration | What it does | How to connect it today |
|---|---|---|
| Looker Studio | Gives you private CSV and JSON addresses for your results, receipts, sources and fixes, which a Looker Studio report reads through a Google Sheet. | Create a data source address |
| NinjaCat | Gives you private CSV and JSON addresses for your results, receipts, sources and fixes for your NinjaCat reports, separate from any other reporting address so each can be revoked on its own. | Create a data source address |
Publishing
| Integration | What it does | How to connect it today |
|---|---|---|
| WordPress | Sends an approved draft from your content workspace to your WordPress site as a draft post. Nothing goes live until you publish it in WordPress. | Connect with an application password |
Commerce and product catalog
| Integration | What it does | How to connect it today |
|---|---|---|
| Shopify | Imports your Shopify products, read only: title, handle, vendor, product type, price and product address, and refreshes them every day. | Upload a Shopify product export |
| Product catalog file | Imports a product catalog from a CSV or JSON file, a Shopify product export included, or from an HTTPS feed address that we read again every day. | Upload a CSV or JSON catalog; Sync from a feed address |
Task management
| Integration | What it does | How to connect it today |
|---|---|---|
| Jira | Sends a fix from your measurement to a Jira project as a task with its observed source, or downloads every fix as a CSV for the Jira importer. | Download fixes for the Jira importer |
| Asana | Sends a fix from your measurement to an Asana project as a task with its observed source, or downloads every fix as a CSV for the Asana importer. | Download fixes for the Asana importer |
| Linear | Sends a fix from your measurement to a Linear team as an issue with its observed source, or downloads every fix as a CSV to import into Linear. | Download fixes for Linear |
Server logs
| Integration | What it does | How to connect it today |
|---|---|---|
| Server logs | Counts visits from AI crawlers in your server logs by day, crawler, page and status, from a Cloudflare Logpush job, a Vercel log drain or an nginx access log, sent to a private address or uploaded as a file. | Create a log drain address; Upload a log file |
Knowledge repositories
| Integration | What it does | How to connect it today |
|---|---|---|
| Notion | Reads the Notion pages you share with your own internal integration, up to 25 at a time, as text with each page's address, last edit and checksum, so you can approve them in the Source library. Read only. | Paste your Notion internal integration secret |
| SharePoint | Reads the text files (Markdown, text, HTML and CSV, up to a megabyte each) in the document library of one SharePoint site your own Microsoft Entra app may read, up to 25 at a time, with each file's address, last edit and checksum. Read only. | Connect with your own Entra app |
| DAM metadata export | Imports the asset list your digital asset manager exports (Bynder, Brandfolder, Cloudinary, Adobe Experience Manager and others export one): title, description, alt text, tags, usage rights, expiry and address, and flags assets whose rights have expired. | Upload a DAM metadata export |
| Document files | Keeps the Markdown, text or HTML documents you upload from any repository (Google Drive, Confluence or a SharePoint export) with a checksum each, up to 100 documents, ready to approve in the Source library. | Upload a document |
Assistant connections: the data and actions each allows
An assistant or tool reaches an account in one of the ways below. Each lists what it may read and the actions it may take. A work owner change or a live question first returns an exact preview and changes nothing until it is confirmed; a free check never starts a paid run. No connection may do any of these: publish anything to your site or any other site; send an email or a message on your behalf; buy, upgrade or change a plan, or change a price; read a connection credential, an API key or a token; read another account, unless your workspace role allows that client. Some reads, such as live question credits, appear only while that feature is on for the account.
| Connection | Reads | Actions | Rate limit |
|---|---|---|---|
| A connector in ChatGPT, Claude or another assistant, after you sign in and choose Allow | List measurements; Results for one measurement; One buyer question; Original answers and cited pages; Fixes for one measurement; Index panel for your category; History by scheduled run; Who is named instead; Pages the engines cited; Controlled lift results; Movements outside sampling noise; Search observed answers; Compare two runs; Client against one competitor; Method and definitions; Get credits; Read connection action history; Read approved source versions; Inspect a saved source version; List buyer-question plans; Read a buyer-question plan | None: it reads only | 60 requests a minute for each connection |
| A connector token you create in the account, read only | List measurements; Results for one measurement; One buyer question; Original answers and cited pages; Fixes for one measurement; Index panel for your category; History by scheduled run; Who is named instead; Pages the engines cited; Controlled lift results; Movements outside sampling noise; Search observed answers; Compare two runs; Client against one competitor; Method and definitions; Get credits; Read connection action history; Read approved source versions; Inspect a saved source version; List buyer-question plans; Read a buyer-question plan | None: it reads only | 60 requests a minute for each connection |
| A connector token with free checks allowed | List measurements; Results for one measurement; One buyer question; Original answers and cited pages; Fixes for one measurement; Index panel for your category; History by scheduled run; Who is named instead; Pages the engines cited; Controlled lift results; Movements outside sampling noise; Search observed answers; Compare two runs; Client against one competitor; Method and definitions; Get credits; Read connection action history; Read approved source versions; Inspect a saved source version; List buyer-question plans; Read a buyer-question plan | Start a free check | 60 requests a minute for each connection |
| A connector token with work owner changes allowed | List measurements; Results for one measurement; One buyer question; Original answers and cited pages; Fixes for one measurement; Index panel for your category; History by scheduled run; Who is named instead; Pages the engines cited; Controlled lift results; Movements outside sampling noise; Search observed answers; Compare two runs; Client against one competitor; Method and definitions; Get credits; Read connection action history; Read approved source versions; Inspect a saved source version; List buyer-question plans; Read a buyer-question plan | Change a work owner | 60 requests a minute for each connection |
| A connector token with live questions allowed | List measurements; Results for one measurement; One buyer question; Original answers and cited pages; Fixes for one measurement; Index panel for your category; History by scheduled run; Who is named instead; Pages the engines cited; Controlled lift results; Movements outside sampling noise; Search observed answers; Compare two runs; Client against one competitor; Method and definitions; Get credits; Read connection action history; Read approved source versions; Inspect a saved source version; List buyer-question plans; Read a buyer-question plan | Ask a live question | 60 requests a minute for each connection |
| An account API key (read only unless you allow live questions or work changes on it) | Every GET route in the API reference for this account and the clients your role allows | Ask a live question, only on a key with live questions allowed; Change a work owner, only on a key with work changes allowed | 60 requests a minute and 5,000 a day for each key |
| A Slack workspace linked with /broadcastwell connect | Your latest scheduled run, with its interval; Your open service work and what waits on you | Delivery decisions, only after an owner or admin allows them on Work from Slack, and only for the Slack members listed there | Slack sends at most one command at a time for each member |
| The public demo connector and demo API key | List measurements; Results for one measurement; One buyer question; Original answers and cited pages; Fixes for one measurement; Index panel for your category; History by scheduled run; Who is named instead; Pages the engines cited; Controlled lift results; Movements outside sampling noise; Search observed answers; Compare two runs; Client against one competitor; Method and definitions | None: it reads only | 30 requests a minute for each network address |
Each account shows the same table beside its own live connections under Delivery, and the API serves it at /api/v1/connector-limits.
How connections treat your data
- Keys, passwords and provider tokens are encrypted before we store them and are never shown again, in the account or in an export.
- We ask for read-only access wherever the provider offers it. Publishing and task connections write only what you send: a draft, never a live page, and one task per fix.
- Disconnecting deletes the connection, its credentials and everything it read, and we ask the provider to revoke our access where it documents a way to.
- Server logs keep counts per day, crawler, page and status. We never keep IP addresses or the log lines themselves.
- Reporting and log drain addresses are shown once and stop working the moment you rotate them or disconnect.
Related: the account API, webhooks and triggers, the export reference for the fields a reporting address serves, and the crawler identity inspector for checking a crawler a log line names.