Weld

06/10/2026

New destination: Apache Iceberg. 🧊

Weld now syncs 300+ sources into Apache Iceberg tables in a bucket you own, through any Iceberg REST catalog, including Amazon S3 Tables, AWS Glue, Polaris and BigLake.

One copy of your data, queryable from Snowflake, Databricks, Athena, Trino or DuckDB.

How it works and how to set it up:

Weld now writes Apache Iceberg tables through any Iceberg REST catalog, including S3 Tables and AWS Glue, so any engine can query one copy of your data.

02/10/2026

New in Weld: a rebuilt ERD viewer for exploring connector schemas.

Search every table and column at once, click a table to see what it joins to, and follow edges that route cleanly around the tables. Here it is on Shopify's 58 tables.

See the demo and try it yourself: https://weld.link/4dMTvMG

01/10/2026

Two different answers to the same question. Numbers that don't match the dashboard. That's usually all it takes for business users to stop trusting an AI agent.

The fix is rarely a better model. It's the context underneath it: skills that tell the agent how to work, an MCP interface, a semantic layer for metric definitions, a context layer for caveats and history, a solid data foundation, and permissions that follow the user.

Here's the agent context stack, layer by layer, and where to start 👇

AI agents give different answers and lose trust. Here's the agent context stack that fixes it: skills, MCP, a semantic layer, context and permissions.

29/09/2026

We're now an official MotherDuck technology partner, and MotherDuck is live as a Weld destination.

That means you can sync all your business data into MotherDuck's serverless DuckDB, model it however you like, and push the results back into the tools your team already uses. Compute is metered per query and scales to zero in between, so you pay for what you actually run.

We've written up three ways teams are using it, including giving AI agents a safe place to query real data.

https://weld.link/4ADLfbx

24/09/2026

🇷🇴 Three days together in Bucharest!

This time, we got together in Romania for our team camp, with colleagues joining from across Europe. It was a great opportunity to step away from the day-to-day, have some fun, and spend quality time together.

We enjoyed some amazing food, explored Bucharest, and most importantly, had three really great days together. 🤝

With a team spread across different countries, getting everyone in the same place is always something special.

Already looking forward to the next one!

11/09/2026

🚀 New connector: Twilio is now live in Weld.

You can now sync your Twilio data directly into your data warehouse and analyze your communications data alongside the rest of your business data.

With the new connector, you can:

→ Sync your Twilio data into your warehouse
→ Keep your data up to date with incremental syncs
→ Combine Twilio data with CRM, billing, product, and other sources
→ Build reporting around communications and customer activity

If Twilio is part of your stack, you can now bring it into your warehouse with Weld.

Read more 👇
https://weld.link/46D29cV

01/09/2026

Weld is now available in the Claude connector directory. 🚀

Find Weld in Claude, click Connect, and you are ready to go. No local setup or config files.

You can ask Claude to work with data from tools like Shopify, HubSpot, and Facebook, using warehouses such as Databricks, Snowflake, and BigQuery.

The connector includes 77 tools across: ingest, transform, orchestrate, and activate.

Everything runs on your existing warehouse and data models, with your current governance in place.

28/08/2026

🚀 New connector: Zoho CRM is now live in Weld.

You can now sync your Zoho CRM data directly into your data warehouse and analyze your sales pipeline alongside the rest of your business data.

With the new connector, you can:
→ Sync 12 Zoho CRM modules, including Deals, Accounts, Contacts, Leads, Calls, Tasks, and Events
→ Keep your data up to date with incremental syncs
→ Combine CRM data with billing, product, marketing, and other sources
→ Build reporting on pipeline, deal flow, rep activity, and account performance

And to help you get started faster, we’ve also released SQL models for:
📊 Deal pipeline
📈 Deal flow by month
☎️ Rep activity
🏢 Account 360

If Zoho CRM is part of your stack, you can now bring it into your warehouse with Weld.

Read more 👇
https://weld.link/4qKW8nB

26/08/2026

Amazon says you sold $412,000 in March.

It deposits $259,000.

Where did the other $153,000 go?

The answer isn’t sitting in one Business Report.

To reconcile the difference, you need settlement data, where fees, refunds, reserves, adjustments, and timing differences all show up on a different schedule.

That’s why many Amazon sellers can tell you revenue per ASIN, but far fewer can tell you profit per ASIN.

We wrote up how to rebuild Seller Central and Vendor Central reporting in your own warehouse, including:

→ Why Amazon’s key revenue numbers don’t match
→ The reporting traps that can quietly inflate your figures
→ How to turn settlement data into a P&L you can reconcile to your payout
→ Why shipped_cogs is what Amazon paid you, while shipped_revenue is what customers paid Amazon

Plus free SQL templates: https://weld.link/4zzmjlj

25/06/2026

🚀 Weld MCP Server: Let Claude Manage Your Data Pipeline

The way we interact with data infrastructure is changing. Instead of jumping between dashboards, logs, and SQL editors, teams can now connect Claude directly to their Weld workspace using MCP (Model Context Protocol) and manage data operations through natural language.

This means Claude can help query your data warehouse, inspect ELT pipelines, investigate issues, and perform operational tasks conversationally. Rather than searching for information across multiple tools, you can simply ask Claude what's happening in your data stack and get answers or take action within the same workflow.

What's particularly interesting is how this shifts software from interfaces built around clicks to interfaces built around conversations. Our latest blog explores what this looks like in practice, including real-world use cases and examples of how teams are integrating Claude into their day-to-day data workflows.

Read the full blog: https://weld.link/4oHXQF4