March 2026

March 8th

Incremental sync for data connectors, embed SDK refinements, scheduling tool improvements, and security fixes.

What’s New

Incremental Sync & Dual Schedules

Data connectors now support incremental sync — each sync records the highest timestamp it saw (e.g., LastModifiedDate), and the next sync fetches only records modified after that point. This replaces the previous full-download-every-time approach for connectors with large datasets.

Cursor-based tracking — QuickBooks filters on MetaData.LastUpdatedTime, HubSpot on hs_lastmodifieddate, Salesforce on LastModifiedDate. Database connectors use whatever timestamp or sequential column you specify (e.g., updated_at)
Dual schedules — a single connector can now have both a full and an incremental schedule. Example: incremental every hour, full refresh every Sunday at 3 AM
Skip sources on full sync — full-sync schedules can skip sources that already have an active incremental schedule, avoiding redundant re-downloads
Source-level skipping — QuickBooks report operations (Profit & Loss, Balance Sheet) don’t have per-record timestamps, so they’re automatically skipped during incremental syncs and only sync during full syncs. The run history shows why each source was skipped
Append, not overwrite — incremental data is appended to the existing dataset rather than replacing it

See the Sync Scheduling guide for per-connector details and setup instructions.

Improvements

Area	Change
Embed SDK	Expired embed sessions now notify the parent window instead of attempting a redirect inside the iframe. “My Account” and “Sign Out” hidden in embed mode. Embed sessions revoked when a user is deleted from the organization
Connector Setup	Source discovery now shows table counts per schema, making it easier to find the right schema in databases with many schemas
Scheduling Tool	Rewritten prompt templates with explicit solver timeouts (30s), concrete examples, and model size limits for more reliable schedule generation
Membership Management	Re-adding a previously removed user now re-establishes their permissions automatically. Removing a user revokes all permissions and cleans up external ID mappings
Database Connectors	Database names containing hyphens (e.g., `manage-qa-redshift-database`) no longer cause parse errors

Bug Fixes

Area	Fix
Row-Level Security	Governed users were seeing unfiltered row counts in the library and unfiltered data in AI analysis prompts. Both now enforce the user’s access policy
Access Policies (API)	V1 API and Embed SDK embed policies were not enforcing `row_filters` — now consistent with dashboard and direct access
Database Columns	Mixed-case column names (e.g., `OrderDate`) were being lowercased during Redshift and Postgres data streaming
Connector Setup	Connector configuration modal was closing when changing form field values
HubSpot	Connector crashed when HubSpot returned zero records for an object type
Connector Syncing	Connectors were auto-processing sources immediately on connection, before the user triggered a sync
Redshift	Table discovery failed on Redshift clusters due to catalog queries using PostgreSQL-specific features (OID lookups) that Redshift doesn’t support

Release Date: March 8, 2026

March 1st

This release brings major platform capabilities to production: a complete Public API, row-level security for governed data access, the Embed SDK on npm, connector scheduling, presentation quality improvements, new analysis tools, and HubSpot enrichment — alongside a series of stability and bug fixes.

What’s New

Public API v1

Querri’s Public API reaches general availability. Authenticate with qk_ API keys and access every major resource programmatically — projects, chats, dashboards, data, embedding, sharing, usage, audit logs, access policies, API key management, users, files, sources, and connectors.

Scoped permissions — each key gets only the scopes it needs
Stripe-style errors — structured error responses with type, code, message, and doc_url
Rate limiting — X-RateLimit-* headers on every response
Audit trail — all API actions are logged with key attribution

See the API Reference and Authentication guide for full details.

Row-Level Security

Access Policies now support full row-level security with shadow execution. Admins define column-based filters (e.g., “Region = East”) and assign them to users or groups. When a governed user runs a query, Querri automatically rewrites it to enforce the policy — no SQL knowledge required.

Shadow execution — policies are evaluated at query time, not at rest, so data stays current
Table filters — admin-defined row restrictions per table
Dashboard filters — interactive filters that viewers can adjust within their allowed scope
Group assignment — apply policies to teams instead of individual users

See the Access Policies guide and Filters guide for setup instructions.

Embed SDK on npm

The Embed SDK is now published as @querri-inc/embed on npm, making it easy to install in any JavaScript project:

npm install @querri-inc/embed

The SDK supports three authentication modes — Share Key for public embeds, Server Token for enterprise/white-label, and Popup Login for internal tools. See the Embed SDK docs for usage details.

Categorize Tool

A new AI-powered Categorize tool discovers hidden themes and topics in your text data — no predefined labels required. Point it at a column of survey responses, support tickets, customer feedback, or any free-text field, and it automatically clusters similar items and generates human-readable category labels.

Discover, don’t define — unlike classification tools that require you to specify categories upfront, Categorize finds the natural groupings in your data
Choose your granularity — pick from Broad (~10–15 themes), Medium (~30–80 topics), or Specific (with sub-categories) after previewing a sample
Confidence scores — every row gets a confidence score so you can filter to high-quality assignments or investigate ambiguous items
Built for downstream analysis — the resulting category columns work directly with charts, aggregations, and dashboards

See the Categorize tool guide for usage tips and examples.

Wrapped Presentation Improvements

Wrapped presentations received a visual quality overhaul:

Responsive charts — Plotly charts now resize correctly across all slide types and screen sizes
Full-viewport mode — presentations fill the entire viewport for a true fullscreen experience
Animation fixes — animations replay correctly on slide transitions and revisits
Download fix — PDF and PowerPoint exports work reliably again

Connector Sync & Scheduling

Data connectors now support scheduled syncs with full run history. Set up recurring syncs on a schedule (hourly, daily, weekly, or custom cron), and Querri keeps your connected data up to date automatically.

Sync scheduling — configure recurring syncs with timezone-aware scheduling
Run history — see past sync runs with per-source detail, duration, and status
Status feedback — live sync progress with grouped batch display

HubSpot Connector Updates

Reduced OAuth scopes — The HubSpot connector no longer requires super admin access. Only 6 core scopes (contacts, companies, deals, owners, tickets, OAuth) are required to connect. Advanced features like sales email, commerce, custom objects, and export are now requested as optional scopes — gracefully skipped if your HubSpot plan doesn’t include them.

Expanded object support — The HubSpot connector now supports 29 object types (up from 20), including dynamic property discovery and custom object support for HubSpot Enterprise users.

Deal enrichment — Deal data is now enriched with human-readable deal stage names and pipeline labels, replacing raw IDs with the labels configured in your HubSpot account.

API Key Management & Audit Log UI

Admins can now create, list, and revoke qk_ API keys directly from the Querri UI. The Audit Log page provides a searchable, filterable view of all security-relevant actions across the organization.

Improvements

Area	Change
API Authentication	Migrated from JWT-based API auth to `qk_` scoped API keys for all public endpoints
Access Policies	Shadow execution model replaces static row filtering for real-time policy enforcement
Embed SDK	Published to npm as `@querri-inc/embed` with ES module support
Wrapped Presentations	Responsive chart sizing, animation replay, and full-viewport mode
HubSpot Connector	Reduced required OAuth scopes from 18 to 6; advanced features use optional scopes. Deal stage and pipeline labels resolve to human-readable names. 9 new object types including custom objects
Connector Scheduling	Recurring sync scheduling with run history, status tracking, and timezone support
Researcher	3x faster batch processing, activity-based timeouts replace hard 30-minute limits, auto-repair of corrupted project data
Dashboard Editor	Image upload and link support in Rich Text Editor widgets

Bug Fixes

Area	Fix
AI Analysis	Fixed intermittent message ordering issues during concurrent AI analysis
Pro Features	Fixed inconsistent pro feature access across the platform
SQL Validation	Fixed false positives when SQL keywords appear inside string literals
Shared Presentations	Fixed 403 errors when viewing shared Wrapped presentations in incognito mode
Analysis Steps	Fixed edge case where running a deleted analysis step could cause errors
Wrapped Download	Fixed broken PDF/PowerPoint export for Wrapped presentations
Slide Animations	Animations now replay correctly on slide transitions and revisits
Column Matching	Fixed column matching failures on files exported from survey tools containing invisible special characters
Researcher	Fixed infinite retry loop that could hang research tasks

Release Date: March 1, 2026

Security & Governance Documentation

A complete documentation section for Querri’s security model, organized around the progressive layers that teams adopt as they grow — 11 pages covering every layer from personal privacy defaults through admin-level controls.

Overview — how Querri keeps your data private by default and makes sharing safe when you’re ready
Sharing — how sharing works and why it’s the foundation for everything else
Access Policies — row-level security without SQL: pick a column, pick values, assign users
Column Security — hide, mask, or restrict sensitive columns while keeping the rest of the table useful
Groups — assign policies to teams instead of managing access user by user
API Keys — qk_ keys for governed, auditable external access that inherits the creator’s permissions
Dashboard Security — how shared dashboards re-execute per viewer’s security context
Filters — admin-defined table filters and interactive dashboard filters
Workspaces — governance boundaries between departments and teams
Audit Log — unified trail of every security-relevant action
Admin Controls — organization-wide settings for sharing, API access, and security defaults

Release Date: February 27, 2026