Reports & Export

The data is good, but the people you have to convince don't live inside Trakkr. They live in inbox, decks, Sheets, a Looker dashboard, or code. So the real question isn't "can I export this," it's: what's the cheapest path from the Reports page to the place the decision actually gets made?

This page covers the five paths. Pick the lowest-effort one that meets your actual freshness requirement, not the most powerful one that meets every requirement you can imagine.

The five exit doors

Scheduled email reports

The Reports page at /reports keeps a sortable timeline of every research run, with a visibility chart and the option to expand any row into the full intelligence briefing. That same briefing, packaged as an email, is what arrives in your stakeholders' inbox on a recurring cadence.

What lands isn't a chart dump. It's an intelligence report: executive summary, hero metrics with seven-day deltas, competitive leaderboard, wins and losses with context, platform breakdowns, strategic recommendations. Designed to be valuable to someone who never opens the dashboard.

Each email carries a shareable report link. On agency plans the whole thing is white-labeled — your logo, your colors, your client portal URL — and the end client never sees Trakkr branding. Full setup: White-Label Portal Overview.

Frequency depends on plan:

One-off CSV and JSON export

For the moments where you just need the data in your hands once — a slide, a snapshot for an analyst, a manual pull for a board pack — there are two paths.

From the Reports page, right-click any row or use the Export dropdown to pull that report (or the whole filtered set) as CSV or JSON. CSV is the row-level summary; JSON is the same plus the nested model_scores and prompt_scores blocks, which is what you want if anything downstream will parse it.

From the export builder at /exports/csv, you choose a mode — Summary (visibility over time, you plus top competitors), Detailed (your rank for every prompt, model, and day), or Raw (all ten ranked brands for every query) — then narrow by date range, models, tags, and columns. A live row count and preview modal let you check the shape before you commit to a 50,000-row download.

Google Sheets sync

The right answer when the team already lives in Sheets and you don't want to be the person emailing a fresh CSV every Monday. Lives at /exports/sheets.

Setup is one-time:

Trakkr then writes — and on every sync, overwrites — five tabs:

The cleanest pattern: one sheet per brand, with your charts and formatted views in separate tabs that read from the five Trakkr ones via formulas. Anything you write directly to a managed tab gets blown away on the next sync.

If verification fails, it's almost always permission. Confirm the service account has Editor (not Viewer) and that the URL points to the intended spreadsheet.

Looker Studio connector

The answer when the request is "we want a real BI dashboard, not a spreadsheet." Lives at /exports/looker-studio, Scale-plan, currently in beta.

It gives you 17 live datasets — Visibility, Citations, Competition, Perception, Crawlers, Live Visitors — each showing up in Looker as a regular data source that recomputes on every dashboard load. Build a dashboard once and it stays fresh.

Setup is two clicks: generate an API key from the page (or use one from /exports/api-keys), then hit the connector button — it opens the Trakkr connector in Looker Studio. Paste the key, pick the brand, you're connected.

What this is genuinely good for is the dashboard you want to send a link to. Looker dashboards live at their own URL, with their own permissions and filter controls — much closer to what an ops or BI team expects than a shared Google Sheet.

REST API and MCP

The developer surfaces. Canonical docs — endpoint reference, code examples in curl/Python/Node, the interactive playground, and the live MCP tool manifest — live at /learn/api.

Reach for the REST API when code is going to consume the data: a warehouse sync, an internal dashboard, a Slack alert, an analytics pipeline. Base URL api.trakkr.ai, bearer-token auth from /exports/api-keys, endpoints covering brands, scores, prompts, citations, competitors, rankings, reports, perception, narratives, crawlers, and more. Webhooks for push-style integrations.

Reach for the MCP server when the consumer is an AI agent — Claude, Cursor, ChatGPT, your own LLM app. MCP exposes the same data as a tool catalog the model can call directly, so you can ask "what's my visibility trend on the pricing-page prompts" and it pulls live data without you writing any API code.

Same engine, two shapes. Pick the one that matches how the consumer thinks.

Three patterns that come up constantly

The weekly client deck. Schedule the weekly report to land in the client's inbox Monday morning, with the shareable report link as the "full picture if you want it." For the deck itself, run a Detailed CSV for the last seven days, pivot in Sheets, paste two or three charts. Under fifteen minutes per client per week.

The exec-leadership dashboard. Set up the Sheets sync at weekly cadence. In a separate tab, build a one-screen scorecard that reads from Summary and Competitor Watchlist: visibility score, seven-day delta, top movers, your rank. Share that tab with leadership. Now the answer to "where are we on AI visibility" is a bookmarked URL.

The BI warehouse pull. REST API on a scheduled job — every six hours is plenty; visibility doesn't move faster than that. Start with /get-scores, /get-reports, and /get-citations, model it however your stack wants. The API is the right tool the moment you'd otherwise be writing a script to scrape your own dashboard.

Plan access

Client portal users see the Reports page and any shared report links the agency has sent them; the export, Sheets, Looker, and API surfaces are agency-side only.

Common questions

Which export gives me the same numbers as the dashboard?

All of them, with one caveat: filters travel. If a date range, model, or tag filter is active on the page you exported from, the export respects it. The most common mismatch is the dashboard sitting on "all time" while the export was scoped to the last 30 days.

Can I schedule a custom report instead of the standard intelligence report?

Not today. Scheduled emails use the intelligence-report template, which is deliberately opinionated about what an executive needs to see. For a different shape, use Sheets sync (raw data, format it yourself) or the REST API (build the email yourself).

Why is my Sheets sync overwriting tabs I edited?

Because that's what it does. The five Trakkr tabs are managed — every sync rewrites them in full. Keep your analysis and formatting in separate tabs that read from the managed ones via formulas.

How fresh is "live" for the Looker Studio connector?

As fresh as the most recent report run. The connector hits the API on every dashboard load, but the underlying reports run daily on Scale and weekly on Growth — so "live" means up-to-date with the latest snapshot, not the literal current minute.

Where do I revoke an API key if it leaks?

/exports/api-keys. Each personal key can be regenerated in place (which kills the old one instantly) or deleted. No grace period — the old key stops working the moment you confirm.