Here’s what we shipped in May 2026.
Below is a running list of some important updates that went live this month. We’ll keep adding to this post as new things ship, so check back often.
May 1
Product Feature: MCP Tables
What changed: Default and last remaining views are now protected from deletion
What we did: Added a guardrail that prevents the MCP from deleting a table’s default view or its only remaining view
Why it matters: No more accidentally leaving a table with no way to access it. The MCP now stops that before it happens
May 1
Product Feature: MCP Date & DateTime
What changed: Full list of date and time format options now available
What we did: Enumerated 30 named date formats and 4 time formats through the MCP
Why it matters: You can now pick the exact date or time display format you need, no guessing or trial and error required
May 1
Product Feature: MCP Page Formatting
What changed: Divider style options are now available through the MCP
What we did: Exposed all five divider styles: Thick, Dashed, Dotted, Wavy, and Paragraph
Why it matters: You can now choose how dividers look on your page, giving you more control over how content is organized visually
May 7
Product Feature: MCP Comments
What changed: You can now resolve and reopen comment threads through the MCP
What we did: Added a dedicated tool for resolving and reopening comment threads
Why it matters: You can close out feedback and reopen it if needed, all without leaving your AI client
May 14
Product Feature: MCP Charts
What changed: All six chart types are now supported as view layouts
What we did: Added bar, line, area, pie, scatter, and bubble charts as options when creating or configuring a table view
Why it matters: Your AI can now build and set up any chart type directly, either as a tab on a table or embedded on another page
May 26
Product Feature: AI Blocks
What changed: Performance fixes and model upgrades for AI Blocks
What we did: Resolved known issues affecting AI Block reliability and updated to the latest models for faster, more accurate output
Why it matters: AI Blocks are built for synthesis at scale. A few things you can now do without the manual effort:
- Summarize at scale: Condense notes, meeting transcripts, or records into a short blurb that auto-refreshes as new data is added
- Surface action items automatically: Pull next steps out of any page or table without reading every row
- Find key themes: Identify patterns across large sets of content without doing the analysis yourself
May 27
Product Feature: MCP Improvements
What changed: A batch of MCP improvements shipped since April, covering charts, comments, view configuration, and more
What we did: See the full launch post for the complete breakdown
Why it matters: More of what you’d do manually in Coda, now doable through your AI client. Want to see it in action? Join the Getting Started with Coda MCP webinar on June 24th.
May 27
Product Feature: AI Columns
What changed: AI Columns are now faster, smarter, and easier to use
What we did: Upgraded to the latest models with batched requests for lower latency, and removed the requirement for @ references. The AI now reads your table structure automatically
Why it matters: Just describe what you want a column to do and Coda handles the rest. A few things you can now do in seconds that used to take an hour:
- Summarize at scale: Condense every row of notes, documents, or records into a short blurb
- Auto-tag and classify: Label hundreds of rows by topic, sentiment, or priority in a single fill
- Surface what matters: Pull specific details out of free-text fields without reading every row
See the full post here.
May 29
Product Feature: Document Extraction
What changed: Document extraction is now generally available
What we did: Enabled AI Chat to parse and reason over file attachments, including PDFs, spreadsheets, and images
Why it matters: You can now ask questions and pull insights directly from files attached to your doc, no manual digging required
May 29
Product Feature: Coda Socket Broker
What changed: Coda Socket Broker is now generally available
What we did: Enabled Packs to make HTTP requests to APIs behind corporate firewalls and private networks
Why it matters: Teams can now use Packs with internal tools and private APIs that were previously out of reach
