Skip to main content

Release Notes

Privacy hardening — ephemeral session identifiers, click IDs no longer stored (August 24, 2026)​

Three privacy improvements are now live for all accounts. Reports, historical data and the API keep working exactly as before, and no action is required.

What changed​

  • Ephemeral session identifiers. The session identifier used to count visits without cookies now expires and regenerates every day, and the key material behind each rotation is destroyed immediately afterwards. Activity from the same device can no longer be linked across days or across sites — by anyone, Sealmetrics included.
  • Click IDs are no longer stored. Ad-platform click identifiers (gclid, fbclid, msclkid and similar) are used only at the moment of the visit to attribute the channel, then discarded — they are not persisted as a field nor inside stored URLs. Campaign attribution works unchanged.
  • Raw User-Agents and IPs removed. The full browser User-Agent string is no longer retained in analytics storage, and edge access logs no longer record visitor IP addresses or User-Agents. Browser, device and country reports are computed in flight and are unaffected.

Heads-up for API and BigQuery users​

If you consume the raw-data endpoints (/stats/conversions/raw, /stats/conversion-items/raw), the clid field no longer appears in the responses. In the BigQuery export, fact_conversions no longer receives click_id: the column remains in existing tables (queries don't break) and is NULL for new rows — and with the click ID out of the aggregation you may see slightly fewer rows, with identical totals. Channel and campaign attribution is unaffected.

Seal AI Private — EU-hosted AI with token packs (July 22, 2026)​

Seal AI Private is now generally available: Sealmetrics' managed AI provider for LENS — processed entirely in the EU (Paris), no prompt retention, and no API key to manage.

What's new​

  • Add-on on Growth, included in Scale and Enterprise. €299/month billed annually upfront (€3,588/year), or €358.80/month on monthly contracts. Not available on Free.
  • 5M tokens per calendar month for the whole organization (chat + automated insights), with email alerts at 80% and 100% consumption.
  • Token packs: extra 5M-token packs at €358.80 each (1–10 per checkout) that never expire — consumed only after the monthly quota.
  • Seal AI Usage screen (Organization → Seal AI Usage, owner/admin): monthly gauge, daily input/output chart, pack balance and history, and per-site / per-user breakdown.
  • BYOK fallback: when the quota runs out, any user can switch the chat to their own Anthropic, OpenAI, Gemini, or DeepSeek key.

Full details: Seal AI Private · LLM Providers.

Sources report — referral traffic now grouped by domain, GA4-style (July 21, 2026)​

The Sources tab now shows referral traffic as source domains (e.g. reddit.com / referrer), unifying historical data under the same convention. The Referrers tab keeps its URL-level detail: one row per referring URL. If you query the API, referral rows now appear in /stats/sources as domains and are excluded from /stats/terms.

Custom Channel Grouping — define your own channels from the dashboard, CSV, or MCP (July 20, 2026)​

You can now personalize how Sealmetrics groups your traffic into marketing channels. The Channels tab in Site settings lets you define your own rules on top of (or overriding) the GA4-style defaults, with a safe drafts → test → publish workflow.

What's new​

  • Rule form for creating custom channels one at a time, with a rule tester ("Test a visit") that uses the exact same engine as the pixel.
  • Override button on default channels — pre-fills the form with the default's patterns so you can redirect any built-in channel to one of your own.
  • CSV import / export (Excel-friendly: , or ; delimiters, RFC 4180 quoting, UTF-8 BOM, CRLF) for bulk edits and version control. Also accepts .json.
  • MCP tools for AI-assisted rule authoring: create_channel_rule, update_channel_rule, delete_channel_rule, import_channel_rules, test_channel_rules. All write tools are draft-only by design — the MCP can never touch a live rule or activate anything; publishing is always a human action from the dashboard.
  • Drafts → test → publish flow: every new rule is born as a draft and never affects traffic until you flip the Live switch. Effective in ~5 min after publishing.

What stays the same​

  • No effect on historical data. Channel classification happens at ingest time; new or edited rules only affect future traffic.
  • Defaults still work out of the box. If you don't create any custom rule, nothing changes for you.

Full walkthrough: Channel Grouping.

Channel classification improvement — more accurate Direct and Referral reporting (July 20, 2026)​

Following feedback from customers who wanted a cleaner mapping between what our tracker captures and the channel groups shown in the dashboard, we've refined the default classification rules for two channels:

  • Direct — traffic with no referrer and no campaign parameters
  • Referral — visits coming from external sites without UTM tagging

The updated rules align our default channel groups with the GA4-style conventions our tracker already uses internally, closing a small gap that had accumulated over successive iterations of the pixel.

Impact on your reports​

Starting July 20, 2026, a portion of the traffic previously reported under Unassigned now appears in Direct or Referral in your channel reports. It's the same traffic you were already receiving — simply labeled more accurately from now on.

A few details worth knowing:

  • The change applies only to new traffic from the effective date onward. Historical rows keep their existing classification (channels are assigned at ingest time).
  • If your account uses custom channel rules, those continue to take priority — this update only affects the built-in defaults.
  • No action is required on your side.

If you have automations, alerts or exports built on the exact volume of Unassigned traffic, this is a good moment to review them.

Bot-blocking algorithm update — fewer false positives (June 2, 2026)​

On Tuesday, June 2, 2026 at 20:00 CET, we deployed an update to our bot-blocking algorithm to stop filtering out legitimate human traffic that was previously being misidentified as bot activity.

What changed​

The algorithm is now more precise at distinguishing real users from automated traffic, reducing false positives without weakening bot detection.

Impact on your reports​

Starting June 2, 2026, you should expect:

  • A slight increase in entries and pageviews, as human visits previously blocked are now counted
  • A more noticeable increase in conversions and microconversions, since affected users were often further along in the funnel
  • No changes to historical data

No action is required on your side.

Attribution: internal hits with UTMs no longer create new sessions (May 25, 2026)​

We improved how Sealmetrics handles internal navigation that carries UTM parameters.

What was happening before​

When a hit on your site contained UTM parameters and the referrer was your own domain (i.e. internal navigation), the system treated it as a new session and attributed that hit to the traffic source declared in the UTMs.

This could inflate session counts and misattribute traffic to campaigns when users were simply navigating across your own pages with UTM-tagged internal links.

What changed​

Starting May 25, 2026, when a hit has UTMs and the referrer is the same domain:

  • The system now prioritizes the referrer over the UTMs
  • The hit is counted as a pageview within the existing session
  • The UTM information is ignored

Impact on your reports​

  • More accurate session counts — no duplicate sessions caused by internal UTM-tagged links
  • Cleaner attribution — campaign sources only reflect genuine external entries
  • Historical data remains unchanged

API: /exports/* and /batch 403 with API keys — fixed (May 2026)​

Fixed a permission check on /exports/* and /batch that required a scope API keys don't carry, returning 403 insufficient_scope for valid keys. The fix is deployed in production. No action required — your existing API key now works on these routes without changes.

See the new API FAQ for the canonical answers on account_id vs site_id, event-level conversions, and /exports/stream examples.

Sealmetrics V2 is Here (February 9, 2026)​

We're thrilled to announce the launch of Sealmetrics V2 — the most significant update since we started.

After months of development and feedback from hundreds of customers, we've rebuilt Sealmetrics from the ground up to deliver:

  • Faster, cleaner dashboard — redesigned for clarity and speed
  • Smarter attribution — understand exactly where your conversions come from
  • Enhanced privacy compliance — ready for GDPR, CNIL, UK PECR, and the upcoming EU Digital Omnibus
  • New API — more powerful, better documented, easier to integrate
  • Improved tracking — lighter script, better SPA support, more accurate data

Sealmetrics V2 is now the default for all accounts. Your data, your settings, and your tracking code continue to work seamlessly.

Thank you to everyone who helped shape this release. We're just getting started.

→ Explore the documentation to discover everything that's new.

Previous Updates​

November 2025​

Update to the Robot User Agents Database (Nov 21, 6:00 UTC)​

We have expanded our internal database of robot-related user agents.
This update adds 158 new user agents, improving the accuracy of our identification and filtering systems.

This enhancement provides:

  • More reliable detection of automated traffic.
  • Higher precision when distinguishing real users from bots.
  • Better performance across our analytics and security workflows.

Check the rest of the documentation to learn how to take advantage of these improvements in your setup.

##Legal Approval for IP-Based Bot Filtering (Nov 18, 20:00 UTC)

After a thorough legal review, we now have official approval to implement IP-based filtering to improve traffic accuracy—without compromising user privacy.

How It Works​

  • When a hit arrives, we check whether the IP is in our bot IP database
  • If it matches → the hit is excluded from your analytics
  • If it doesn’t match → we do not store the IP, we simply register it as human traffic

Privacy First​

  • We do not retain or track IPs that aren't bots
  • No identification or storage of human IPs
  • Full compliance with privacy regulations

This ensures:

  • Better protection of your analytics data
  • No compromise on user privacy
  • No personal data is ever stored or exposed

It’s a technical improvement — but also a principle: You can have precision without violating privacy. You can have control without crossing the line.

Facebook Traffic Classification Fix (NEW – Nov 14, 19:00 UTC)​

We've improved how Sealmetrics classifies Facebook traffic to ensure higher accuracy between organic and paid sources.

What Was Happening Before​

A minor issue caused some visits containing the fbclid parameter (but without UTM tags) to be incorrectly categorized as facebook-ads.

This affected a small portion of traffic, but could distort organic vs. paid Facebook reporting.

What Changed​

As of Wednesday, November 17 at 19:00 UTC:

  • Traffic containing only fbclid and no UTM parameters
    → is now classified as Facebook Organic (correct behavior)

  • Only traffic containing proper UTM campaign parameters
    → is classified as Facebook Ads

Impact on Your Reports​

  • Improved accuracy in distinguishing paid vs. organic Facebook traffic
  • Expect a slight increase in organic Facebook traffic from this timestamp onward
  • Historical data remains unchanged to preserve consistency

Performance Improvements (Nov 10, 22:00 UTC)​

2.5x Faster Data Processing​

We've significantly optimized our processing infrastructure by increasing core capacity. Your data now processes 2.5 times faster, meaning real-time dashboard updates and instant report generation.

Benefits:

  • Faster dashboard loading times
  • Real-time data updates
  • Instant report generation
  • Improved overall platform responsiveness

Facebook Ads Attribution Update (Nov 13, 19:00 UTC)​

Breaking Change

As of this release, the fbclid parameter is no longer used to identify Facebook Ads traffic due to its lack of reliability and consistency.

What this means for you:

  • Facebook Ads traffic now requires properly configured UTM tags for accurate attribution
  • Ensure your campaigns include parameters such as:
    • utm_source=facebook
    • utm_medium=paid
    • (or equivalent)
  • This ensures consistent, accurate attribution and full control over your data

Recommendation: Review and update your Facebook Ads campaigns with correct UTMs to maintain full visibility.

tip

Need help configuring UTMs? Check our Facebook Ads Tracking Guide

Migration Guide​

If you're currently relying on fbclid for Facebook Ads attribution, follow these steps:

  1. Audit your campaigns: Check all active Facebook Ads campaigns
  2. Add UTM parameters: Configure UTMs in your Facebook Ads Manager
  3. Test: Verify traffic attribution inside Sealmetrics
  4. Monitor: Ensure continuity in your reporting

Need assistance? Contact our support team