Integrations
The hub at /integrate, where you connect Trakkr to your CMS, your edge, your analytics, and the tools your team already works in.
On this page
Integrations is the hub at /integrate, where you connect Trakkr to the rest of your stack. It lists every service Trakkr supports, shows what you have connected and whether it is healthy, and gives every connection the same connect, test, and disconnect flow. The old /exports and /connect pages redirect here.
Open it when a feature looks empty and you suspect the connection behind it, when a card shows a warning, or when you want a new destination for alerts and tickets. Browsing the catalog is open to any account; connecting needs a paid plan. Client portal users cannot reach the hub.
Data in, work out
Every integration does one of two jobs.
Bring data in. Connect a system that knows something Trakkr does not. Google Analytics knows which visitors arrived from an AI assistant. Cloudflare, Vercel, and Netlify know which AI bots crawled which pages. Your CMS is what Trakkr reads before it proposes a fix.
Send work out. Connect a system your team already lives in, so findings land there instead of in another tab. Slack gets the alert, Linear gets the ticket, Google Sheets gets the rows.
The six categories
The hub's left rail is All plus six categories. Each one shows a count of how many connections you have configured in it.
| Category | What lives there |
|---|---|
| Your website | WordPress, Shopify, Webflow, GitHub. Connect a site to publish fixes |
| AI traffic | AI Crawler Tracking, AI Pages, Google Search Console, Google Analytics |
| Advertising | OpenAI Ads. Connect ad accounts to act on what you find |
| Alerts & tasks | Slack, Microsoft Teams, Discord, Linear, Jira, GitHub Issues, Asana, Trello, Notion, Zapier, Make |
| Export | CSV Export, Google Sheets, Looker Studio (Scale), Airtable (Soon) |
| Developer | Webhooks, REST API, MCP Server |
Above the categories sits Your connections, one band listing everything you have configured with its live state. Problems sort to the top. The header repeats it as a single line: how many are configured, how many are healthy, and how many need attention.
Connect an integration
The shape is the same almost everywhere.
- 1.Open the hub and find the card. Cards you cannot use yet read Soon or Unavailable instead of Connect.
- 2.Authorize. WordPress, GitHub, Vercel, Netlify, Slack, Linear, Jira, Notion, Asana, Trello, and Teams hand off to the provider's consent screen. Cloudflare, Webflow, and Looker Studio take a token you paste in. Shopify takes a myshopify domain plus a client ID and secret.
- 3.Pick the destination. Slack asks for a channel, Linear asks for a team, Sheets asks for a spreadsheet. You can change it later.
- 4.Test it. Webhook-style connections have a Test connection action that sends a sample event.
Some cards open in place. WordPress opens a drawer with its three capabilities side by side (Publishing, Crawler tracking, AI Pages), each with its own Enable, because they are separate connections to the same site.
What is scoped to a brand, and what is not
This trips people up, so it is worth stating plainly.
Per brand. Site connections, Sheets sync, crawler tracking, Google Analytics, Search Console, OpenAI Ads, and the work-item accounts behind Slack, Linear, Jira, GitHub Issues, Notion, Asana, and Trello. Connecting one on Nike does not connect it on Adidas.
Per account. Outgoing webhooks and the cards built on them (Zapier, Make, Discord, Teams, Webhooks), plus your REST API key. These belong to you, not to the brand you were viewing.
The directory
Grouped by what a service does for you. A service that does more than one thing is listed where it spends most of its time.
Bring data in
| Integration | What it gives Trakkr |
|---|---|
| Google Analytics | Sessions and conversions attributed to AI referrers, shown in Traffic > Visitors |
| Google Search Console | Google search performance next to AI visibility |
| AI Crawler Tracking | Server-side AI bot hits. Cloudflare uses an API token, Vercel and Netlify use OAuth, WordPress uses the Trakkr plugin, and self-hosted stacks (Next.js, Node, Nginx, CloudFront, Akamai, Fastly) post to a webhook |
| Hosted CMS platforms | Webflow, HubSpot, and Squarespace expose no readable server logs, so their crawler tracking routes through Cloudflare. Shopify and Wix do not allow a proxy, so crawler tracking is not available there. See Install crawler tracking |
| WordPress, Shopify, Webflow, GitHub | Read access to the pages Trakkr audits and writes to. See Sites |
| AI Pages | Serves AI-optimized versions of your pages to crawlers at the edge |
Send work out
| Integration | What Trakkr can send to it |
|---|---|
| Slack | Approval requests, notifications, action buttons |
| Microsoft Teams | Automation alerts to a Teams channel |
| Discord | Automation alerts to a server channel |
| Linear, Jira, GitHub Issues, Asana, Trello, Notion | An issue, task, card, or page created from a Trakkr finding |
| Google Sheets | Five tabs of visibility data, refreshed manually, daily, or weekly |
| Looker Studio | Seventeen live datasets for BI dashboards. Scale plan. See Looker Studio |
| CSV Export | Scoped downloads of rankings and citations, no setup |
| Webhooks | Any event payload to any HTTPS endpoint, with no auth, Bearer, Basic, or an API key header |
| Zapier, Make | Trakkr as a step in your existing automations. See Zapier |
| REST API | Build it yourself. Scale plan. See the API reference |
| MCP Server | Your Trakkr brand as tools for ChatGPT, Claude, Cursor, or Codex. Any paid plan |
| Gmail | Send Trakkr reports from your own address, where the mailbox feature is enabled |
Tickets and messages are configured at /integrate/work-items, on three tabs: Connections for accounts, Destinations for the channels, boards, and projects you send to, and Routing for the rules deciding which finding goes where.
Security and permissions
Saved secrets stay server-side. Site credentials are encrypted with AES-256-GCM before they are stored. Integration responses return safe connection details, never the saved token, webhook URL, or secret. Secret fields are write-only once saved.
Narrow scopes. Trakkr asks for the smallest permission set each job needs. Slack asks for chat:write, chat:write.public, and channels:read, nothing about user identity. Linear asks for read, write, and issue creation. Google Sheets sync uses a service account you share one spreadsheet with, so Trakkr never sees the rest of your Drive.
Hosted MCP clients. A connected assistant can ask Trakkr for the categories the MCP tools expose: brands, prompts, visibility, citations, competitors, reports, crawler and Action data, automations, webhooks, Reddit monitoring, and related account metadata. Some tools are read-only, others change state when you explicitly ask. Hosted clients receive short-lived connector tokens, not a raw API key, and you can revoke access from Settings → Developer.
ChatGPT. Trakkr is published in the ChatGPT directory. Open the Trakkr listing, choose Install plugin, then sign in to Trakkr and approve access. The published flow does not need Developer mode or an MCP connect token. Start with one of these:
- "Show my latest AI visibility score and the biggest change this month."
- "Find my top citation gaps and recommend the first three pages to improve."
- "Compare my brand across ChatGPT, Perplexity, Gemini, and Claude."
When a plan lapses. The hub does not hide connections you can no longer use. A Connection cleanup section lists them so you can revoke stored access at any time, entitlement or not.
Checking what was sent. Open a webhook-family connection and switch to its delivery history: delivered or failed, the response status, and the payload. Its row menu also carries Test connection and Remove integration.
Common questions
Do I have to connect anything to use Trakkr?
No. Prompts, scores, research, citations, competitors, perception, and the dashboard all work with no integration. Integrations unlock the features that need an outside system: publishing through Sites, crawler tracking, visitor attribution, alerts, and exports.
My tool is not in the directory. What are my options?
Webhooks send JSON to any HTTPS endpoint, which covers most internal systems. Zapier and Make bridge Trakkr to thousands of apps without code. MCP exposes your data to any assistant that speaks Model Context Protocol. Past that, the REST API on Scale.
Can one Slack workspace serve several brands?
Yes. Work-item accounts are stored per brand, so you connect the workspace once per brand and point each one at a different channel. The underlying Slack install can be shared.
Why does a card say Unavailable rather than Connect?
Site cards check with the backend first. If a platform cannot currently accept a new connection, the card says Unavailable instead of sending you into a dead end, though one already connected stays clickable. Soon means the integration is not built yet.
What happens to data already sent if I disconnect?
It stays. Past Slack messages, Linear issues, Sheets rows, and applied site fixes all persist in the destination. Disconnecting only stops new sends.
Why does a connection show a warning instead of an error?
Trakkr separates "broken" from "needs attention" from "could not check". An expired token needs attention and offers a reconnect. If the status check itself fails, the hub says the status is unavailable rather than calling a working connection disconnected.