Welcome to the last release of 2025!
We've been building toward AI-assisted review workflows all year, and this month brings it together: improved AI summaries to help during development (and before opening a PR) and a real-time team workspace to discuss and approve it. Redshift support is live, and we just wrapped our team offsite in Taiwan planning what's next.
Let's get into it.
We've made several upgrades to how AI summaries work:
Data validation is a team sport, but the tooling hasn't always reflected that. Slack discussions. Anonymous approvals. Stakeholders can't access the review because they don't have a GitHub account. We've fixed all of this.
Shared Sessions: Your whole team now works in the same live environment. One session spins up and stays live for 6 hours, everyone connects instantly to the same workspace. No waiting for your own instance. You're collaborating in real-time, in the same space, looking at the same checks.
Inline Check Comments: Every check now supports threaded comments. Reviewers and developers can discuss specific validations directly in context — not in a Slack thread three clicks away. Conversations persist across sessions and stay attached to the checks they reference.
Transparent Approval Workflow: Approvals now show exactly who approved each check and when. No more anonymous sign-offs. You get a clear audit trail for compliance, debugging, and accountability.
Frictionless Stakeholder Access: Bring domain experts (Finance, Compliance, Analytics) into the review process with just a Recce Cloud account. No GitHub or GitLab accounts required. If they need to review data, they should be in Recce.
You asked, we shipped. Recce Cloud now supports Amazon Redshift natively. Redshift teams get the same diffs, lineage checks, and PR validations that our Snowflake, BigQuery, and Postgres users rely on. Available now in General Availability.
This was a big year for Recce. Here's what we shipped:
Recce Cloud: We launched an entirely new product. What started as open-source tooling evolved into a full cloud platform for data validation and review.
The Pivot to Agents: We bet early on AI-assisted review workflows. The AI Summary, MCP integration, and our "human-in-the-loop" philosophy all came together this year. We're not just building tools anymore — we're building a review agent.
The Team Grew: We closed our pre-seed round and welcomed 4 new team members. We're now a team of 10, fully dedicated to making data validation effortless.
Data Renegades Podcast: We launched a podcast that interviews pioneers in the data space. Episodes 1-5 are now live! Check out conversations with Benn Stancil (founder of Mode), Simon Wilison (founder of Django & Datasette), Max Beauchemin (creator of Airflow & Superset), Micah Wylde (creator of Arroyo). Listen on Apple Podcasts, Spotify, or stream directly here.
Community Moments: Shoutout to the team at Viadukt for partnering on a case study and being incredible early adopters. And thank you to Matteo who presented about Recce organically at a meetup this month, that meant more than you know.
What's Next in 2026: Our focus is simple: more value with less effort. We want data validation to be so easy that there's no excuse not to do it. That means helping you not just at the PR stage, but during development. Faster feedback loops, better defaults, and fewer clicks. Time and annoyance are the main reasons teams skip validation today. We're going to fix that.
The full Recce team (all 10 of us!) gathered in Taiwan this month for a 3-day offsite in a hot springs town on the coast, just east of Taipei.
We ran a year-in-review, mapped out the 2025 roadmap, and held an internal hackathon. Some highlights from the hack projects: early prototypes for a chatbot agent to help run reviews, better tooling for dbt Fusion compatibility, and improvements to our own internal data workflows.
We also ran an unconference-style session where everyone pitched topics and voted on what to discuss.
The biggest takeaway? Relentless focus on making Recce easier to use. We know we've been a power tool, sometimes a complicated one. In 2026, we're prioritizing intuitive defaults and a smoother learning curve so more teams can get value on day one.
With the launch of MCP and BigQuery, we are planning our roadmap for Q1. We want to hear what you need next.
If you can spare 30 minutes to chat, we’ll send you a $25 gift card as a thank you.
Your feedback drives everything we build. Connect with us: