Roundtable Summary: Building Made Easy with Coda AI

[Ruggy_Joesten1]

Hey all!

Thanks to those of you who were able to attend today’s roundtable, giving a sneak peek into what’s next for building with Coda AI. Our guest of honor @Bharat_Batra1 was able to walk us through the following:

Coda MCP

• Per Bharat, Coda MCP functions like a “USB stick” for Coda data; portable knowledge that can plug into any AI tool.

• Currently in private beta >> let us know if you are interested in accessing, and we can share details.

• Enables external AI tools (Cursor, Claude Code, ChatGPT) to read from and write to Coda.

• Particularly powerful for technical teams using AI coding assistants to create documentation directly from codebases.

• Can read Coda comments and suggest fixes based on team feedback.

• Coming soon: ChatGPT and Claude Web integration.

New Coda AI Features

• Complete redesign of Coda AI with iterative building capabilities.

• Can create complex tables, filtered views, and relationships with natural language requests.

• Understands Coda-specific patterns and best practices (statuses, priorities, filters, etc.).

• Provides guided suggestions for next steps in workflow setup.

• Works inline with document context for better accuracy.

• Will eventually support automation creation from chat interactions.

Technical Improvements

• Both MCP and Coda AI use new frameworks (not the existing Coda API) for better performance.

• Context window optimizations with dynamic tool discovery (ToolGuide feature).

• Client-side execution where possible for faster response times.

Philosophy

• Create a powerful AI companion that understands workflows, works where users do, and helps them move faster.

• Eliminate “work behind the work” to reduce time spent on schema decisions, formatting, and setup.

• Make Coda expertise accessible to all users of all skill levels.

Some questions we were able to answer live on the call:

1. @Cristian_Nichifor — Inline (in-row) AI vs sidebar

   • Question / ask: Support inline AI entry points (e.g., in a row/detail view) instead of or in addition to the sidebar for better context, discoverability, history of prompts, guardrails.

   • Answer (summary): Bharat said it’s great feedback; they’re thinking about it and exploring multiple locations (table, formula editor, automations, rows). Bharat acknowledged its value and is investigating, but no firm implementation details or timeline were provided.

2. @Agile_Dynamics — What other MCP tools are you working on? (buttons / automations)

   • Question: Request for automation capabilities via MCP (need buttons / push actions) and what tools are planned.

   • Answer: Bharat explained context-window constraints and ToolGuide; first three tools coming: add comments (already added), add images to pages, and push buttons (for automation triggers). More tools will follow as models/context improve. Coda AI will get tools faster than MCP,

3. @Agile_Dynamics (follow-up) — Will Go SDK agents (Superhuman Go SDK) have MCP access?

   • Question: Will agents built with the Go SDK have access to MCP features (same capabilities)?

   • Answer: Yes. It works internally; the team is figuring out how to expose it to customers (auth flow questions remain). Functionality exists internally; customer-facing access and auth details are TBD.

4. @Hugo_Assuncao1 — AI hard-coded dates vs. Coda formulas

   • Question: The agent hard-coded a “next week” date instead of using a Coda formula, can it generate proper Coda formulas?

   • Answer: It can use Coda formulas but sometimes hard-codes values; team plans to improve prompting/eval sets and UI to show what the agent did so users can detect/fix mistakes.

5. @Hugo_Assuncao1 — Delay between document changes and AI availability (API latency)?

   • Question: Is there a delay (as with the Coda API) between doc changes and AI seeing them?

   • Answer: Coda MCP and the new Coda AI do not use the regular Coda API; they use a new framework with client-side execution where possible for near-instant availability; some operations remain server-side.

6. @Greg_Villeneuve — How to guide what context Coda AI responds to (page/table/doc/workspace)?

   • Question: Can you guide the AI to look at specific pages/tables/docs? Does it know the rest of the docs?

   • Answer: Yes. First version: you can provide a URL (or ID); default is your current selection/page. Team will continue tuning context levels and personalization.

7. @Greg_Villeneuve (follow-up) — Other connectors / external tools planned?

   • Question: What connectors are planned and how will external tools be accessed?

   • Answer: Two approaches: 1) Go platform connectors (many Coda packs → Go connectors), and 2) investigate connecting other MCPs into Coda AI chat (not day one).

8. @Kayla_Taras2 — Workspace-level context (consolidating where a user is tagged across docs)

   • Question: Can workspace-level context be used to consolidate a user’s scattered tags/rows across many docs into one place (e.g., “find everywhere I’m tagged” and create a consolidated doc)?

   • Answer: Bharat acknowledged the idea and said it makes sense; called out as a good use case for MCP (daily prep agent) and something they’d like to enable.

Updates to unanswered questions from yesterday

1. @Sarah_Heyborne1 Will Coda MCP work with Copilot?

Answer: Yes, as long as copilot supports connecting to remote MCP servers like other providers. Reading this release, it seems like they do:

2. @Greg_Villeneuve What’s most effective for guiding coding agents (context efficiency, safety, AGENTS.md, skills, etc.)?

Answer: For context efficiency, the best results tend to come from the model providers themselves. Coda supports dynamic tool discovery and relies on agents (e.g., Claude) to use it effectively.

Skills / AGENTS.md / guardrails are powerful for personalization. For example, you can:

• Instruct agents to use specific Coda templates

• Require approval before editing certain docs

• Enforce doc-specific safety rules

3. @Greg_Villeneuve Beta / launch timeline for refreshed Coda AI?

Answer: Aiming for a private beta in April.

4. @Melanie_Teh Add a “Chat” tab at the row level (alongside Comments & Activity)

Answer: Love it. Chat history is already planned, and the team will think through row-level implementations as well.

5. @Pablo_DV Have you considered CLI tools that are more token-efficient than MCP?

Answer: Not a CLI tool specifically, but a Completion API has been considered, which would allow using the agent from anywhere. It has not been officially prioritized yet.

6. @Cristian_Nichifor Would a GA Coda MCP include agents that already “know” Coda (vs. needing to understand comments dynamically)?

Answer: Skills also consume context. The broader industry trend is toward dynamic tool loading (similar to how skills work). For now, the plan is to continue improving tool discovery and descriptions rather than baking in fixed “knowledge.”

7. Will Coda MCP work in Google Antigravity (without custom setup)?

Answer: Yes, this already works today. Documentation is available here.

8. @Cristian_Nichifor Can we get chat history at the account/workspace level to build a shared knowledge base?

Answer: Chat history is planned at the user/account level. The team is also exploring ways to make knowledge shareable to prompts.

We greatly value your feedback and insights. The generosity and passion from our community fuel so much of our decision-making, and sessions like these remind us how fortunate we are to work alongside such engaged and thoughtful folks. Thank you for showing up, asking great questions, and helping us build better tools together. We’re grateful for each of you and can’t thank you enough for being part of the Coda community.

Our next roundtable happens next Thursday at 11am PT. We hope you’ll join us!

[Jordan_Axani]

Love these recaps, @Ruggy_Joesten1!

[Josh_Mullins]

Hi! I’m interested in gaining access to the CODA MCP. Will the capabilities available on the MCP server eventually be accessible via the API? For instance, we would like to sync our CODA docs with a centralized database to support other AI functionalities. We’ve considered moving away from CODA because this feature isn’t currently available. While MCP is a step in the right direction, long-term API access would better meet our needs.

[Michael_Warren]

echoing what @Josh_Mullins said — very interesting in exploring the MCP server as we are slowly moving more and more of our business off of Coda.

[Pashmina_De_Shon]

I’d love to get in on the Coda MCP functions beta.

Also is there anywhere to review what you all discussed?

[Jeb_Dasteel]

Great update. I’d love to be part of the private beta, please.

[Luis_Matos]

I am interested in Coda MCP functions like a “USB stick” for Coda data; portable knowledge that can plug into any AI tool.

Thank you

[Daniel_Stieber]

Awesome! I‘m very interested in trying and joining the beta. Already tried out an open source version that uses the api but that sounds way more promising!!

[Avishek_Kedia1]

interested in the MCP. Please add me.

[Thomas_Schulz1]

Thanks for sharing @Ruggy_Joesten1 very useful recap and kudos to the participants

[lars_hofstetter_10TIMES_io]

Hi Ruggy, sounds exciting, I’m also interested in the Beta. Best, Lars

[Ed_Liveikis]

Hi @Ruggy_Joesten1 , we’re interested in the MCP if possible to add.

[Mike_Clark]

Exciting - we’ve been hoping MCP was coming soon! Would love to be part of this beta.

[Ruggy_Joesten1]

Love seeing the excitement around the discussion, and enthusiasm for the MCP beta test! I DMed each of you who requested access

[Kevin_Tinoco]

Hi @Ruggy_Joesten1, I’m interested in the MCP access if I can still get in.

[Ruggy_Joesten1]

@Ruggy_Joesten1

Updates to unanswered questions from yesterday

All other Dory questions answered! Take a look at the edited post to find yours.

[Dan_Feder]

Hi @Ruggy_Joesten1 –thanks for sharing these notes, and I’m sorry I missed the roundtable. This sounds very exciting! I saw your note about letting you know if we’re interested in access to the private MCP beta–I filled out the waitlist form when it initially opened, but would love if I could get access! I’ve been using some 3rd party MCP tools in the meantime, but they aren’t ideal.

[Ruggy_Joesten1]

Copy! Let me nudge now.

[Bill_French]

@Ruggy_Joesten1

a sneak peek into what’s next for building with Coda A

Really appreciated the great recap from the roundtable!

It’s fantastic to see Coda laying out a robust vision for where AI is heading—especially the focus on iterative building, schema support, and breaking AI out of the typical “sidebar” jail.

Having spent the last several months working deeply with the private beta of the Coda MCP server via Antigravity and Claude Code, I wanted to jump in and offer a bit of a “reality check.” The roundtable summary significantly understates what is already possible today when you combine Coda MCP with a competent agentic framework. And by “competent”, I mean “additive contexts and rules”.

The notes mention that the first three MCP tools coming are things like pushing buttons, adding images, and commenting. But in the current beta environment, my agents have access to over 30 distinct tools made available in Coda MCP. I’m already orchestrating full CRUD operations across Documents, Pages, Tables, Columns, Rows, and even dynamically building and executing CFL Formulas.

Through an integration framework I’ve been developing called mcpOS, I’ve learned that I do not have to wait for Coda AI to figure out schemas or relational logic [eventually]. By combining the Coda MCP server with rigorous, deterministic workflows, orchestrating these complex structural builds has been possible for months now. In fact, mcpOS has directly anticipated and remedied several of the challenges mentioned in the roundtable including, but not limited to these:

1. Schema Hallucinations and Data Pollution: The digest noted that AI sometimes struggles with Coda-specific patterns and hard-codes values. mcpOS remedies this via Schema Defense (/coda-schema). Before writing any data, the agent dynamically audits the target table’s columns and formats, ensuring it never attempts to write to a non-existent column or insert mismatched data types.
2. Context Window Limits and Blindspots: Answering questions about “the rest of the docs” or workspace-level context is a known challenge. mcpOS remedies this using a Shadow Context Protocol (/coda-shadow), which actively maps document structure (pages, canvases, formulas) to gain deterministic context, bypassing the limitations and blind spots of standard RAG approaches that are likely used inside Coda and almost certainly the cause of these issues.
3. Complex Logic and Duplication: The native AI is still working toward understanding complex relational data. mcpOS handles this today with Smart Upserts (/coda-upsert). It programmatically handles the logic needed to safely add or update records without creating duplicates, acting much more like a traditional database integration than a simple chatbot.

As several of my Community articles show, a deep dependence on Coda’s internal AI features comes with some downsides. I’m more of an “agentic pureist” who believes that agentic platforms (which Coda has chosen to be) should implement safety and defenses for everyday use cases. But we cannot—and should not—expect the AI tooling to do everything because with everything that is possible comes the expense of agility.

Coda’s philosophy of eliminating the “work behind the work” is absolutely spot-on, and validates my assertions in several MCP-related articles. As predicted here, here, here, and here—Makers will soon be operating with a new definition of making Coda solutions.

It’s important that the community knows that, by leveraging Coda’s MCP through rigorous workflows that anyone can create in Markdown files with tools like Antigravity, Cursor, Windsurf, or custom agentic setups, you have the license to functionally operate months ahead of the native roadmap laid out in this digest. Here’s another prediction:

Because of the agentic abilities to utilize MCP with agent-native rules and visual reasoning of Coda itself, which can be designed and implemented seamlessly, Makers will forever have slightly more control than Coda developers.

Exciting times ahead!

[Elise_Tarabiscode.com]

Hello @Ruggy_Joesten1

Thank you for this summary. I’m delighted to see all these changes coming, and can’t wait to try them out!

I signed up for the beta, but I haven’t received access. Could you give me access?

Thank you very much.