Superhuman Docs (Coda AI).

Coda continues under a new name

Superhuman Docs lets you create and edit material directly in a shared document instead of copying responses from a separate chat. It combines text, structured tables, formulas, buttons, and connections to other systems. This helps with project summaries, requests, and working reference documents: AI output stays beside the source records. The official announcement explicitly confirms Coda’s transition to Superhuman Docs and the preservation of existing documents. Listing both brands separately would count one product twice.

Text and data in one workspace

Docs AI can help with drafts, summaries, tables, and understanding team context. AI columns handle repeated processing of records, while blocks summarize pages and tables. Keep a clear distinction between source facts and calculated or generated content. For example, a ticket status from a table must not be replaced with a guess from free text.

The new brand also announced AI Views and Superhuman Databases, but the official announcement labels them closed beta. They cannot be promised to every user as generally available features. Base your workflow on features actually enabled in your account.

Example: prepare a decision log

You have your own anonymized notes from three meetings on one project. Ask AI to create a table with “decision — rationale — owner — source,” including only explicitly agreed decisions. Put proposals in a separate section. If no owner is specified, leave the field blank rather than assigning someone merely because their name appears nearby.

Open each row’s source and check that a discussion has not become a commitment. Add standard status fields and make the document understandable to someone who missed the meetings. For ongoing use, one well-maintained log is more useful than a long automatically generated summary. This is an educational scenario, rather than our test of Docs AI accuracy.

Integrations and access

Official Docs MCP lets compatible AI clients read and modify documents. The REST API covers documents, pages, and table rows; it does not establish a universal endpoint for every interactive AI action. An integration should check token permissions, the delay before changes apply, and confirmation that data was actually written.

Pricing depends on the plan and Doc Maker role; check AI access and quota details in the current workspace. A free document does not prove ongoing free AI access. Before connecting external data, set member permissions and ensure summaries do not disclose private table contents to unauthorized people.

A decision log benefits from links back to original notes. Wording changes should still let readers understand what was said and what the team approved later.

Suitable tasks

• Create a decision log from notes
• Compile a summary of the project table
• Process similar records using an AI column
• Connect a document to an external AI client